THIRD
EDITION
Pump
Characteristics
and Applications
Michael Volk
Pump
Characteristics
and Applications
THIRD
EDITION
Pump
Characteristics
and Applications
THIRD
EDITION
Michael Volk
Boca Raton London New York
CRC Press is an imprint of the
Taylor & Francis Group, an informa business
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2014 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works
Version Date: 20130507
International Standard Book Number-13: 978-1-4665-6309-4 (eBook - PDF)
This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been
made to publish reliable data and information, but the author and publisher cannot assume responsibility for the
validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the
copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to
publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let
us know so we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted,
or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written
permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com
(http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers,
MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety
of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment
has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
Contents
Preface to the Third Edition.................................................................................xv
Preface to the Second Edition............................................................................ xvii
Preface to the First Edition.................................................................................. xix
Acknowledgments............................................................................................. xxiii
Author................................................................................................................... xxv
1. Introduction to Pumps.................................................................................... 1
I. What Is a Pump?.................................................................................... 1
II. Why Increase a Liquid’s Pressure?...................................................... 2
III. Pressure and Head.................................................................................3
IV. Classification of Pumps.........................................................................4
A. Principle of Energy Addition.......................................................4
1. Kinetic................................................................................... 4
2. Positive Displacement......................................................... 4
B. How Energy Addition Is Accomplished....................................6
C. Geometry Used..............................................................................6
V. How Centrifugal Pumps Work............................................................6
VI. PD Pumps.............................................................................................. 12
A. General.......................................................................................... 12
B. When to Choose a PD Pump..................................................... 12
C. Major Types of PD Pumps......................................................... 15
1. Sliding Vane Pump............................................................ 19
2. Sinusoidal Rotor Pump..................................................... 20
3. Flexible Impeller Pump.................................................... 20
4. Flexible Tube (Peristaltic) Pump...................................... 21
5. Progressing Cavity Pump................................................22
6. External Gear Pump.......................................................... 23
7. Internal Gear Pump.......................................................... 25
8. Rotary Lobe Pump............................................................ 26
9. Circumferential Piston, Bi-Wing Lobe Pumps................ 27
10. Multiple-Screw Pump....................................................... 28
11. Piston Pump....................................................................... 29
12. Plunger Pump.................................................................... 31
13. Diaphragm Pump.............................................................. 32
14. Miniature PD Pumps........................................................ 35
2. Hydraulics, Selection, and Curves............................................................. 39
I. Overview............................................................................................... 39
II. Pump Capacity..................................................................................... 41
III. Total Head.............................................................................................42
© 2008 Taylor & Francis Group, LLC
v
vi
Contents
IV.
V.
VI.
VII.
VIII.
IX.
X.
XI.
XII.
XIII.
A. Static Head...................................................................................43
B. Friction Head............................................................................... 46
C. Pressure Head.............................................................................. 57
D. Velocity Head............................................................................... 58
Performance Curve.............................................................................. 60
Horsepower and Efficiency................................................................ 68
A. Hydraulic Losses......................................................................... 70
B. Volumetric Losses....................................................................... 70
C. Mechanical Losses...................................................................... 70
D. Disk Friction Losses.................................................................... 70
NPSH and Cavitation.......................................................................... 76
A. Cavitation and NPSH Defined.................................................. 76
1. NPSHa.................................................................................. 81
2. NPSHr..................................................................................84
B. Calculating NPSHa: Examples................................................... 85
C. Remedies for Cavitation............................................................. 86
D. More NPSHa Examples............................................................... 89
E. Safe Margin NPSHa versus NPSHr........................................... 92
F. NPSH for Reciprocating Pumps............................................... 96
Specific Speed and Suction Specific Speed...................................... 97
Affinity Laws...................................................................................... 103
System Head Curves.......................................................................... 107
Parallel Operation.............................................................................. 116
Series Operation................................................................................. 122
Oversizing Pumps............................................................................. 126
Pump Speed Selection....................................................................... 128
A. Suction Specific Speed.............................................................. 129
B. Shape of Pump Performance Curves..................................... 129
C. Maximum Attainable Efficiency............................................. 129
D. Speeds Offered by Manufacturers.......................................... 133
E. Prior Experience........................................................................ 133
3. Special Hydraulic Considerations............................................................ 135
I. Overview............................................................................................. 135
II. Viscosity.............................................................................................. 135
III. Software to Size Pumps and Systems............................................. 150
A. General........................................................................................ 150
B. Value of Piping Design Software............................................ 151
C. Evaluating Fluid Flow Software............................................. 152
D. Building the System Model...................................................... 153
1. Copy Command............................................................... 154
2. Customize Symbols......................................................... 154
3. CAD Drawing Features.................................................. 154
4. Naming Items................................................................... 154
5. Displaying Results........................................................... 155
© 2008 Taylor & Francis Group, LLC
vii
Contents
6. The Look of the Piping Schematic................................. 155
Calculating the System Operation.......................................... 155
1. Sizing Pipe Lines............................................................. 156
2. Calculating Speed............................................................ 156
3. Showing Problem Areas................................................. 156
4. Equipment Selection....................................................... 156
5. Alternate System Operational Modes........................... 157
F. Communicating the Results.................................................... 157
1. Viewing Results within the Program........................... 157
2. Incorporating User-Defined Limits.............................. 157
3. Selecting the Results to Display.................................... 157
4. Plotting the Piping Schematic........................................ 158
5. Exporting the Results...................................................... 158
6. Sharing Results with Others.......................................... 158
7. Sharing Results Using a Viewer Program.................... 158
G. Conclusion.................................................................................. 158
H. List of Software Vendors.......................................................... 159
IV. Piping Layout...................................................................................... 159
V. Sump Design....................................................................................... 165
VI. Field Testing........................................................................................ 166
A. General........................................................................................ 166
B. Measuring Flow......................................................................... 167
1. Magnetic Flowmeter....................................................... 168
2. Mass Flowmeter............................................................... 168
3. Nozzle................................................................................ 168
4. Orifice Plate...................................................................... 168
5. Paddle Wheel.................................................................... 169
6. Pitot Tube.......................................................................... 169
7. Segmental Wedge............................................................ 169
8. Turbine Meter................................................................... 169
9. Ultrasonic Flowmeter...................................................... 169
10. Venturi............................................................................... 170
11. Volumetric Measurement............................................... 170
12. Vortex Flowmeter............................................................ 170
C. Measuring TH............................................................................ 171
D. Measuring Power...................................................................... 173
E. Measuring NPSH...................................................................... 173
E.
4. Centrifugal Pump Types and Applications........................................... 175
I. Overview............................................................................................. 175
II. Impellers.............................................................................................. 176
A. Open versus Closed Impellers................................................ 176
B. Single versus Double Suction.................................................. 182
C. Suction Specific Speed.............................................................. 183
D. Axial Thrust and Thrust Balancing....................................... 185
© 2008 Taylor & Francis Group, LLC
viii
Contents
III.
IV.
V.
VI.
VII.
VIII.
IX.
X.
XI.
XII.
XIII.
XIV.
XV.
XVI.
E. Filing Impeller Vane Tips......................................................... 187
F. Solids Handling Impellers....................................................... 189
End Suction Pumps............................................................................ 190
A. Close-Coupled Pumps.............................................................. 190
B. Frame-Mounted Pumps........................................................... 193
Inline Pumps...................................................................................... 195
Self-Priming Centrifugal Pumps..................................................... 197
Split-Case Double Suction Pumps................................................... 199
Multistage Pumps.............................................................................. 203
A. General........................................................................................ 203
B. Axially Split-Case Pumps........................................................ 203
C. Radially Split-Case Pumps...................................................... 207
Vertical Column Pumps.................................................................... 208
Submersible Pumps........................................................................... 213
Slurry Pumps...................................................................................... 215
Vertical Turbine Pumps..................................................................... 218
Axial Flow Pumps............................................................................. 226
Regenerative Turbine Pumps........................................................... 227
Pump Specifications and Standards................................................ 228
A. General........................................................................................ 228
1. Liquid Properties............................................................. 229
2. Hydraulic Conditions...................................................... 229
3. Installation Details.......................................................... 229
B. ANSI............................................................................................ 230
C. API............................................................................................... 232
D. ISO............................................................................................... 233
Couplings............................................................................................234
Electric Motors.................................................................................... 240
A. Glossary of Frequently Occurring Motor Terms.................. 240
1. Amps................................................................................. 240
2. Code Letter....................................................................... 241
3. Design Letter.................................................................... 241
4. Efficiency........................................................................... 242
5. Frame Size......................................................................... 242
6. Frequency......................................................................... 242
7. Full-Load Speed............................................................... 242
8. High Inertial Load........................................................... 242
9. Insulation Class................................................................ 242
10. Load Types........................................................................ 242
11. Phase.................................................................................. 243
12. Poles................................................................................... 243
13. Power Factor..................................................................... 243
14. Service Factor................................................................... 244
15. Slip..................................................................................... 244
16. Synchronous Speed......................................................... 244
© 2008 Taylor & Francis Group, LLC
ix
Contents
B.
C.
D.
E.
F.
G.
H.
I.
17. Temperature..................................................................... 244
18. Time Rating...................................................................... 244
19. Voltage............................................................................... 245
Motor Enclosures...................................................................... 245
1. Open Drip-Proof.............................................................. 245
2. Totally Enclosed Fan Cooled.......................................... 245
3. Totally Enclosed Air Over.............................................. 246
4. Totally Enclosed Nonventilated.................................... 246
5. Hazardous Location........................................................ 246
Service Factor............................................................................. 246
Insulation Classes..................................................................... 247
Motor Frame Size...................................................................... 247
1. Historical Perspective..................................................... 247
2. Rerating and Temperature............................................. 250
3. Motor Frame Dimensions............................................... 251
4. Fractional Horsepower Motors...................................... 251
5. Integral Horsepower Motors.......................................... 251
6. Frame Designation Variations....................................... 251
Single-Phase Motors................................................................. 257
Motors Operating on Variable Frequency Drives................ 261
NEMA Locked Rotor Code...................................................... 262
Amps, Watts, Power Factor, and Efficiency........................... 263
1. Introduction...................................................................... 263
2. Power Factor..................................................................... 263
3. Efficiency........................................................................... 264
4. Amperes............................................................................ 265
5. Summary.......................................................................... 265
5. Sealing Systems and Sealless Pumps...................................................... 267
I. Overview............................................................................................. 267
II. O-Rings................................................................................................ 267
A. What Is an O-Ring?................................................................... 268
B. Basic Principle of the O-Ring Seal.......................................... 268
C. The Function of the O-Ring..................................................... 268
D. Static and Dynamic O-Ring Sealing Applications............... 270
E. Other Common O-Ring Seal Configurations........................ 270
F. Limitations of O-Ring Use....................................................... 271
III. Stuffing Box and Packing Assembly............................................... 272
A. Stuffing Box................................................................................ 273
B. Stuffing Box Bushing................................................................ 273
C. Packing Rings............................................................................ 273
D. Packing Gland............................................................................ 274
E. Lantern Ring.............................................................................. 275
IV. Mechanical Seals................................................................................ 276
© 2008 Taylor & Francis Group, LLC
x
Contents
A.
Mechanical Seal Advantages................................................... 276
1. Lower Mechanical Losses.............................................. 276
2. Less Sleeve Wear.............................................................. 276
3. Zero or Minimal Leakage............................................... 276
4. Reduced Maintenance..................................................... 276
5. Seal Higher Pressures..................................................... 276
B. How Mechanical Seals Work................................................... 277
C. Types of Mechanical Seals....................................................... 280
1. Single, Inside Seals.......................................................... 280
2. Single, Outside Seals....................................................... 282
3. Single, Balanced Seals..................................................... 283
4. Dual Seals......................................................................... 283
5. Gas Lubricated Noncontacting Seals............................ 286
6. Seal Piping Plans............................................................. 288
V. Sealless Pumps................................................................................... 292
A. General........................................................................................ 292
B. Magnetic Drive Pumps............................................................. 293
1. Bearings in the Pumped Liquid.................................... 295
2. Dry Running.................................................................... 296
3. Inefficiency....................................................................... 296
4. Temperature..................................................................... 296
5. Viscosity............................................................................ 297
C. Canned Motor Pumps.............................................................. 297
1. Fewer Bearings................................................................. 299
2. More Compact.................................................................. 299
3. Double Containment....................................................... 299
4. Lower First Cost............................................................... 299
6. Energy Conservation and Life-Cycle Costs........................................... 301
I. Overview............................................................................................. 301
II. Choosing the Most Efficient Pump.................................................. 302
III. Operating with Minimal Energy..................................................... 307
IV. Variable-Speed Pumping Systems...................................................308
V. Pump Life-Cycle Costs...................................................................... 325
A. Improving Pump System Performance: An Overlooked
Opportunity?............................................................................. 326
B. What Is Life-Cycle Cost?.......................................................... 327
C. Why Should Organizations Care about Life-Cycle Cost?......327
D. Getting Started.......................................................................... 328
E. Life-Cycle Cost Analysis.......................................................... 328
1. Cic—Initial Investment Costs......................................... 330
2. Cin—Installation and Commissioning (Start-Up)
Costs.................................................................................. 330
3. Ce—Energy Costs............................................................ 331
4. Co—Operation Costs....................................................... 332
© 2008 Taylor & Francis Group, LLC
xi
Contents
5.
6.
7.
F.
G.
H.
I.
J.
Cm—Maintenance and Repair Costs............................. 332
Cs—Downtime and Loss of Production Costs............334
Cenv—Environmental Costs, Including Disposal
of Parts and Contamination from Pumped Liquid....334
8. Cd—Decommissioning/Disposal Costs,
Including Restoration of the Local Environment.........334
Total Life-Cycle Costs............................................................... 335
Pumping System Design.......................................................... 335
Methods for Analyzing Existing Pumping Systems............ 339
Example: Pumping System with a Problem Control Valve.... 341
For More Information...............................................................344
1. About the Hydraulic Institute........................................346
2. About Europump.............................................................346
3. About the U.S. Department of Energy’s Office of
Industrial Technologies..................................................346
7. Special Pump-Related Topics................................................................... 347
I. Overview............................................................................................. 347
II. Variable-Speed Systems....................................................................348
III. Sealless Pumps...................................................................................348
IV. Corrosion............................................................................................. 349
A. Galvanic or Two-Metal Corrosion.......................................... 351
B. Uniform or General Corrosion................................................ 351
C. Pitting Corrosion....................................................................... 352
D. Intergranular Corrosion........................................................... 353
E. Erosion Corrosion...................................................................... 353
F. Stress Corrosion......................................................................... 354
G. Crevice Corrosion...................................................................... 354
H. Graphitization or Dezincification Corrosion.........................354
V. Nonmetallic Pumps........................................................................... 355
VI. Materials Used for O-Rings in Pumps............................................ 357
A. General........................................................................................ 357
1. Polymer............................................................................. 357
2. Rubber............................................................................... 357
3. Elastomer.......................................................................... 357
4. Compound........................................................................ 358
B. Eight Basic O-Ring Elastomers................................................ 358
1. Nitrile (NBR, Buna N)..................................................... 358
2. Neoprene........................................................................... 359
3. Ethylene Propylene (EP, EPR, and EPDM)................... 359
4. Fluorocarbon (FKM, Viton, and Kalrez)....................... 359
5. Butyl................................................................................... 360
6. Polyacrylate....................................................................... 360
7. Silicone.............................................................................. 360
8. Fluorosilicone................................................................... 362
© 2008 Taylor & Francis Group, LLC
xii
Contents
VII.
VIII.
IX.
X.
High-Speed Pumps............................................................................ 363
Bearings and Bearing Lubrication................................................... 365
Precision Alignment Techniques..................................................... 366
Software to Size Pumps and Systems............................................. 367
8. Installation, Operation, Maintenance, and Repair............................... 369
I. Overview............................................................................................. 369
II. Installation, Alignment, and Startup.............................................. 369
A. General........................................................................................ 369
B. Installation Checklist................................................................ 370
1. Tag and Lock Out............................................................ 370
2. Check Impeller Setting................................................... 370
3. Install Packing or Seal..................................................... 371
4. Mount Bedplate, Pump, and Motor............................... 371
5. Check Rough Alignment................................................ 371
6. Place Grout in Bedplate................................................... 373
7. Check Alignment............................................................. 373
8. Flush System Piping........................................................ 373
9. Connect Piping to Pump................................................ 374
10. Check Alignment............................................................. 375
11. Turn Pump by Hand....................................................... 375
12. Wire and Jog Motor......................................................... 375
13. Connect Coupling............................................................ 376
14. Check Shaft Runout......................................................... 376
15. Check Valve and Vent Positions.................................... 376
16. Check Lubrication/Cooling Systems............................ 376
17. Prime Pump if Necessary............................................... 376
18. Check Alignment............................................................. 377
19. Check System Components Downstream.................... 377
20. Start and Run Pump........................................................ 377
21. Stop Pump and Check Alignment................................ 378
22. Drill and Dowel Pump to Base...................................... 378
23. Run Benchmark Tests...................................................... 378
III. Operation............................................................................................ 378
A. General........................................................................................ 378
B. Minimum Flow.......................................................................... 379
1. Temperature Rise............................................................. 379
2. Radial Bearing Loads...................................................... 380
3. Axial Thrust..................................................................... 380
4. Prerotation........................................................................ 380
5. Recirculation.................................................................... 381
6. Settling of Solids.............................................................. 382
7. Noise and Vibration........................................................ 382
8. Power Savings, Motor Load........................................... 383
© 2008 Taylor & Francis Group, LLC
xiii
Contents
C.
D.
Preferred Operating Range...................................................... 383
Ten Ways to Prevent Low Flow Damage in Pumps.............384
1. Continuous Bypass.......................................................... 385
2. Multicomponent Control Valve System........................ 385
3. Variable Frequency Drive............................................... 386
4. Automatic Recirculation Control Valve........................ 387
5. Relief Valve....................................................................... 388
6. Pressure Sensor................................................................ 388
7. Ammeter........................................................................... 389
8. Power Monitor.................................................................. 389
9. Vibration Sensor............................................................... 389
10. Temperature Sensor........................................................ 390
IV. Maintenance....................................................................................... 390
A. Regular Maintenance............................................................... 390
1. Lubrication........................................................................ 390
2. Packing.............................................................................. 391
3. Seals................................................................................... 392
B. Preventive Maintenance........................................................... 392
1. Regular Lubrication......................................................... 392
2. Rechecking Alignment................................................... 392
3. Rebalance Rotating Element.......................................... 392
4. Monitoring Benchmarks................................................. 393
C. Benchmarks................................................................................ 393
1. Hydraulic Performance................................................... 393
2. Temperature..................................................................... 393
3. Vibration........................................................................... 394
V. Troubleshooting................................................................................. 399
VI. Repair...................................................................................................400
A. General........................................................................................400
B. Repair Tips................................................................................. 401
1. Document the Disassembly........................................... 401
2. Analyze Disassembled Pump........................................ 402
3. Bearing Replacement...................................................... 402
4. Wear Ring Replacement................................................. 403
5. Guidelines for Fits and Clearances...............................404
6. Always Replace Consumables.......................................404
7. Balance Impellers and Couplings................................. 405
8. Check Runout of Assembled Pump.............................. 405
9. Tag Lubrication Status.................................................... 405
10. Cover Openings Prior to Shipment............................... 405
9. Case Studies.................................................................................................. 407
I. Introduction........................................................................................ 407
II. Case Studies........................................................................................ 407
© 2008 Taylor & Francis Group, LLC
xiv
Contents
A.
B.
C.
D.
E.
F.
G.
H.
I.
The Case of the Oversized Pump........................................... 407
1. Background...................................................................... 407
2. Analysis of the Problem..................................................408
3. Solutions and Lessons Learned..................................... 409
The Case of the Unreliable Refrigerant Pump...................... 412
1. Background...................................................................... 412
2. Analysis of the Problem.................................................. 413
3. Solutions and Lessons Learned..................................... 417
The Case of the Vibrating Vertical Turbine Pump............... 418
1. Background...................................................................... 418
2. Analysis of the Problem.................................................. 419
3. Solutions and Lessons Learned..................................... 420
The Case of Too Many Pumps................................................. 421
1. Background...................................................................... 421
2. Analysis of the Problem..................................................422
3. Solutions and Lessons Learned..................................... 426
The Case of Too Few Pumps.................................................... 427
1. Background...................................................................... 427
2. Analysis of the Problem.................................................. 431
3. Solutions and Lessons Learned..................................... 431
The Case of the Underperforming Pump.............................. 433
1. Background...................................................................... 433
2. Analysis of the Problem.................................................. 435
3. Solutions and Lessons Learned..................................... 435
The Case of the Problematic Variable Speed Pump............. 437
1. Background...................................................................... 437
2. Analysis of the Problem.................................................. 438
3. Solutions and Lessons Learned.....................................440
The Case of the Shoe-Horned Wastewater Pumps...............440
1. Background......................................................................440
2. Analysis of the Problem..................................................442
3. Solutions and Lessons Learned.....................................444
The Case of the High Suction Specific Speed Pump............445
1. Background......................................................................445
2. Analysis of the Problem..................................................445
3. Solutions and Lessons Learned.....................................446
Appendix A: Major Suppliers of Pumps in the United States by
Product Type........................................................................................................ 449
Appendix B: Conversion Formulae................................................................. 461
References............................................................................................................ 473
© 2008 Taylor & Francis Group, LLC
Preface to the Third Edition
This third edition of Pump Characteristics and Applications includes two significant improvements. First, more than 150 images are presented in color
for the first time. Based on feedback from participants in short courses in
pumps that have used this book as the text, these color images should greatly
improve the ability of readers from all backgrounds to understand the details
in many cutaway and cross sectional images of pumps, seals, and other
components. Similarly, color-coding permits a clear distinction between the
many types of pump performance curves that are presented throughout the
book, such as head–capacity, horsepower, efficiency, NPSH, and system head
curves. Also included are new images of the latest generation of pumps and
other components.
The second focus of this edition is a new chapter on pump case studies. Students in Volk pump training classes indicate that case studies are
an important learning tool for people who work with pumps. The new
Chapter 9 describes in some detail a series of typical pump field problems
and their solutions. Each case study includes background on the pumprelated problem, an analysis of the problem, and the resulting solutions and
lessons learned.
In addition to the above, the entire book was updated to reflect the latest
thinking on pumps. A number of helpful new sections were added, such
as the ten steps to determine total head that are summarized in Chapter 2
and the mechanical seal piping plans that are discussed and illustrated in
Chapter 5.
The first two editions of the book included a demo CD of the PIPE-FLO
Professional piping design and analysis software. With this third edition we
have switched to a downloadable demo of the software. PIPE-FLO and other
piping design and analysis software programs are discussed in some depth
in Chapter 3. Use of a software tool to design or analyze a piping system
should be considered if the system is complex (e.g., has multiple branches or
loops, includes multiple pumps or a single pump at multiple speeds, or uses
a fluid with properties much different from water). Using a software tool
to size pipes and determine pump total head allows the piping and pumps
to be matched more accurately to the expected demands of the system, and
helps to keep the pump from being oversized, thus saving energy and reducing pump maintenance costs. It also gives the system designer or operator a
much better understanding of how the pump responds to changes in levels,
flows, pressures, and valve positions. For more information on PIPE-FLO
and to download a demo of this software, go to the following link: http://
www.volkassociates.com/pipeflo.html.
© 2008 Taylor & Francis Group, LLC
xv
xvi
Preface to the Third Edition
In closing, I would like to invite readers who are interested in attending
a more in-depth training course on pumps, either for continuing education
units or simply to enhance their understanding of pumps and systems, to go
the link below for information on short courses in pumps for engineers and
technicians that I periodically offer at a variety of venues, including on-site
at company offices (http://www.volkassociates.com/seminars.html).
© 2008 Taylor & Francis Group, LLC
Preface to the Second Edition
Thankfully, the laws of physics have not changed since the first edition of
this book was written in 1996. Therefore, virtually everything about pump
selection, sizing, system analysis, and other aspects of pump hydraulics
remains unchanged from the first edition. There have, however, been a number of innovations in the world of pumps, which are introduced in this second edition. This edition also expands the material on many components of
typical pump installations that were only briefly covered in the first edition,
if at all. Some of the most important new or expanded topics covered in this
second edition include
• Chapter 1—Several new types of positive displacement (PD) pumps
are introduced, while the information on other types of PD pumps
has been expanded.
• Chapter 2—Important new topics in this chapter include NPSH analysis for closed systems, expansion of the discussion on NPSH margin, and system head curve development for existing systems and
for parallel pumping systems.
• Chapter 3—In the world of software, nine years is an eternity, and so
the entire section of this chapter covering software used to design
and analyze pump piping systems has been completely rewritten.
• Chapter 4—Entire new sections of this chapter have been added to
provide in-depth coverage of two very important and relevant topics: pump couplings and electric motors. Additionally, several types
of centrifugal pumps that were not included in the first edition are
covered in this chapter.
• Chapter 5—This chapter has an entire new section on O-rings
used in pumps, as well as additional information about sealless
pumps.
• Chapter 6—Two major additions to the book are included in this
chapter. The first is an in-depth discussion of variable-frequency
drives. Second, this chapter includes a section covering pump
life-cycle cost, an innovative approach to the study of the cost of
pumping equipment that looks way beyond the capital cost of the
pump.
• Chapter 7—This chapter has added in-depth discussion of metallic
corrosion in pumps, as well as discourse on elastomers commonly
used in pumps for sealing components.
© 2008 Taylor & Francis Group, LLC
xvii
xviii
Preface to the Second Edition
• Chapter 8—New topics covered in this chapter include ten methods
to prevent low flow damage in pumps, and a much more detailed
discussion of vibration, including a detailed vibration troubleshooting chart.
Thanks to my colleagues in the pump field who provided input for this
second edition or who reviewed particular sections of it. Finally, I wish to
thank my daughter Sarah, who typed major portions of the new material for
this edition.
© 2008 Taylor & Francis Group, LLC
Preface to the First Edition
This book is a practical introduction to the characteristics and applications
of pumps, with a primary focus on centrifugal pumps. Pumps are among
the oldest machines still in use and, after electric motors, are probably the
most widely used machines today in commercial and industrial activities.
Despite the broad use of pumps, this subject is covered only briefly in many
engineering curricula. Furthermore, companies that use pumps are often
unable to provide their engineers, operators, mechanics, and supervisors the
kind of training in pump application, selection, and operation that this vital
equipment merits.
The purpose of this book is to give engineers and technicians a general
understanding of pumps and to provide the tools to allow them to properly select, size, operate, and maintain pumps. There are numerous books
on the market about pumps, but most of them are very technical, and are
mainly design oriented, or else are directed to a specific niche market. I have
attempted to provide practical information on pumps and systems to readers
with all levels of experience, without getting so immersed in design details
as to overwhelm the reader.
This book begins with the basics of pump and system hydraulics, working
gradually to more complex concepts. The topics are covered in a clear and
concise manner and are accompanied by examples along the way. Anyone
reading the material, regardless of education and experience with pumps,
will be able to achieve a better understanding of pump characteristics and
applications.
Although it is not possible to cover pump hydraulics without getting into
some mathematics, this book covers the subject without resorting to differential equations and other high level mathematics that most people forgot
right after school. For the reader who is interested in a more complex or
sophisticated approach to particular topics or who wants additional information in a given area, references are made to other sources that provide a
more analytical approach.
A theme that is repeated throughout this book is that all aspects of
pumps—from system design, to pump selection, to piping design, to installation, to operation—are interrelated. Lack of attention to the sizing of a
pump or improper design of the piping system can cause future problems
with pump maintenance and operation. Even the most precisely sized pump
will not perform properly if its installation and maintenance are not performed carefully. A better understanding of how these issues are related will
help to solve problems or to prevent them from occurring in the first place.
© 2008 Taylor & Francis Group, LLC
xix
xx
Preface to the First Edition
In addition to a thorough treatment of the fundamentals, this book also
provides information on the current state of the art of various technologies in the pump field. Variable speed pumping systems, sealless pumps,
gas-lubricated noncontacting mechanical seals, and nonmetallic pumps are
examples of recent technological trends in the pump industry that are introduced in this book. Computer software for system design and pump selection is previewed in Chapter 3. This is another example of a powerful new
technology related to pumps that is covered in this book.
Because the book focuses on pump applications and characteristics, rather
than on design, it is intended for a broader audience than typical books about
pumps. The readership for this book includes the following:
• Engineers—This book has broad appeal to mechanical, civil, chemical, industrial, and electrical engineers. Any engineer whose job it is
to design or modify systems; select, specify, purchase, or sell pumps;
or oversee operation, testing, or maintenance of pumping equipment will find this book very helpful.
• Engineering supervisors—Because they have broad responsibility for
overseeing the design and operation of pumps and pump systems,
engineering supervisors will benefit from the integrated systems
approach provided in this book.
• Plant operators—Employees of plants that utilize pumps are required
to oversee the operation of the pumps, and often their maintenance,
troubleshooting, and repair. A better understanding of hydraulics
and applications will help these people do a better job of operating
their pumps most efficiently while reducing maintenance costs and
downtime.
• Maintenance technicians—Maintenance personnel and their supervisors can do a much better job of installing, maintaining, troubleshooting, and repairing pumps if they have a better understanding
of how pumps are applied and operated in a system.
• Engineering students—The “real-world” problems that are presented
in this book demonstrate to students that a pump is more than a
“black box.” Many university engineering departments are expanding their technology program to better prepare students for jobs in
industry. This book can make an important contribution to a program in industrial machinery.
The formulae used in this book will generally be stated in U.S. Customary
System (USCS) units, the system most widely used by the pump industry in
the United States. Appendix B at the end of this book provides simple conversion formulae from USCS to SI (metric) units. The most common terms
mentioned in this book will be stated in both units.
© 2008 Taylor & Francis Group, LLC
Preface to the First Edition
xxi
I wish to thank my colleagues in the pump field who reviewed various
sections of this book or who assisted in obtaining materials and illustrations.
I am especially grateful to my friends Jim Johnston, Paul Lahr, and Buster
League, who reviewed the entire manuscript and provided me with valuable
feedback. Final thanks go to my wife, Jody Lerner, for her word processing
and editorial skills as well as for her patience and encouragement.
© 2008 Taylor & Francis Group, LLC
Acknowledgments
Thanks to my colleagues in the pump field, Greg Case and Ray Hardee, who
reviewed the new sections of this edition. Special thanks to Jeff Sines and
his employer, Engineered Software, for supplying the resources and talent to
generate many of the color images and pump curves, as well as his helpful
review. And of course, I would not have the dedication to the development of
my career in pumps if it were not for the constant support of my wife, Jody
Lerner, and my daughters Monica and Sarah.
© 2008 Taylor & Francis Group, LLC
xxiii
Author
Michael W. Volk, PE, is the president of Volk & Associates, Inc., Oakland,
California, www.volkassociates.com, a consulting company specializing
in pumps and pump systems. Volk’s services include pump training seminars; pump equipment evaluation, troubleshooting, and field testing; expert
witness for pump litigation; witnessing of pump shop tests; pump market
research; and acquisition and divestiture consultation and brokerage. A
member of the American Society of Mechanical Engineers (ASME) and a registered professional engineer, Volk received his BS degree (1973) in mechanical engineering from the University of Illinois, Urbana, and his MS degree
(1976) in mechanical engineering and his MS degree (1980) in management
science from the University of Southern California, Los Angeles. He may be
contacted at mike@volkassociates.com.
© 2008 Taylor & Francis Group, LLC
xxv
1
Introduction to Pumps
I. What Is a Pump?
Simply stated, a pump is a machine used to move liquid through a piping system and to raise the pressure of the liquid. A pump can be further
defined as a machine that uses several energy transformations to increase
the pressure of a liquid. The centrifugal pump shown in Figure 1.1 illustrates this definition. The energy input into the pump is typically the
energy source used to power the driver. Most commonly, this is electricity
used to power an electric motor. Alternative forms of energy used to power
the driver include high-pressure steam to drive a steam turbine, fuel oil to
power a diesel engine, high-pressure hydraulic fluid to power a hydraulic
motor, and compressed air to drive an air motor. Regardless of the driver
type for a centrifugal pump, the input energy is converted in the driver to a
rotating mechanical energy, consisting of the driver output shaft, operating
at a certain speed, and transmitting a certain torque. The power transmitted from the driver to the pump is a function of the rotating speed times
the torque.
The remaining energy transformations take place inside the pump itself.
The rotating pump shaft is attached to the pump impeller (see Figure
1.4). The rotating impeller causes the liquid that has entered the pump to
increase in velocity. This is the second energy transformation in the pump,
where the input power is used to raise the kinetic energy of the liquid.
Kinetic energy is a function of mass and velocity. Raising a liquid’s velocity
increases its kinetic energy.
After the liquid leaves the impeller, but before exiting the pump, the
final transformation of energy occurs in a diffusion process. An expansion
of the flow area causes the liquid’s velocity to decrease to more than when
it entered the pump, but well below its maximum velocity at the impeller tip. This diffusion transforms some of the velocity energy to pressure
energy.
1
2
Pump Characteristics and Applications
Liquid out:
High pressure
Fuel
source
energy in:
Air, steam, electricity,
hydraulic fluid, or diesel
engine oil
Driver
Pump
Rotating
mechanical
energy
Liquid in:
Low
pressure
FIGURE 1.1
A centrifugal pump uses several energy transformations to raise the pressure of a liquid.
II. Why Increase a Liquid’s Pressure?
There are actually three distinct reasons for raising the pressure of a liquid
with a pump, plus another related factor:
1. Static elevation. A liquid’s pressure must be increased to raise the liquid from one elevation to a higher elevation. This might be necessary, for example, to move liquid from one floor of a building to a
higher floor, or to pump liquid up a hill.
2. Friction. It is necessary to increase the pressure of a liquid to move
the liquid through a piping system and overcome frictional losses.
Liquid moving through a system of pipes, valves, and fittings experiences frictional losses along the way. These losses vary with the
geometry and material of the pipe, valves, and fittings, with the viscosity and density of the liquid, and with the flow rate.
3. Pressure. In some systems, it is necessary to increase the pressure of
the liquid for process reasons. In addition to moving the liquid over
changes in elevation and through a piping system, the pressure of a
liquid must often be increased to move the liquid into a pressurized
vessel, such as a boiler or fractionating tower, or into a pressurized
pipeline, or it may be necessary to overcome a vacuum in the supply
vessel.
4. Velocity. There is another factor to be considered here, namely that
not all of the velocity energy in a pump is converted to potential or
pressure energy. The outlet or discharge connection of most pumps
is smaller than the inlet or suction connection. Because liquids are,
practically speaking, incompressible, the velocity of the liquid leaving the pump is higher than that entering the pump. This velocity
head may need to be taken into account (depending on the point of
reference) when computing pump total head to meet system requirements. This is discussed further in Chapter 2, Section III.
3
Introduction to Pumps
III. Pressure and Head
It is important to understand the relationship between pressure and head.
Most plant engineers and those involved in operations tend to speak of the
pressure of the liquid at various points in the process. Pressure is measured
in pounds per square inch (psi, sometimes simply called “pounds”) if U.S.
Customary System (USCS) units are used. In SI (metric) units, the equivalent
units for pressure are kilopascal (kPa), bar, or kilograms per square centimeter (kg/cm2), whereas the equivalent units for head are meters (m). Most
readers should be familiar with the difference between gauge pressure and
absolute pressure. Absolute pressure is gauge pressure plus atmospheric
pressure. Atmospheric pressure is 14.7 psi (1 bar) at sea level.
In the study of pump hydraulics, it is important to realize that any pressure
expressed in psi (kPa) is equivalent to a static column of liquid expressed in
feet (meters) of head. This is not meant to imply that pressure and head are
interchangeable terms, because conceptually head is a specific energy term
and pressure is a force applied to an area. However, the units used in hydrodynamics for specific energy are ft-lb/lbf, which, in the gravitational field of
the earth where the acceleration of gravity is 32.2 ft/s2, can be numerically
reduced to (feet of) head. With this understanding (and because most pumping applications occur under the earth’s gravitational influence), the terms
will be used interchangeably in this book.
The equivalence between pressure and head is illustrated in Figure 1.2,
where the pressure in psi read on a gauge located at the bottom of a column
of liquid is related to the height of the column in feet and to the specific gravity
of the liquid by the following formula (in USCS units):
Pressure (psi) =
Water
(SG = 1.0)
Head (ft) × SG
2.31
100 ft
= 100 ft × 1.0 = 43 psi
2.31
43 psi
FIGURE 1.2
Pressure (in psi) is equivalent to a vertical column of liquid with a certain specific gravity.
4
Pump Characteristics and Applications
psi =
feet × SG
2.31
(1.1)
where psi is the pounds per square inch and SG is the specific gravity.
Specific gravity is the weight of a given volume of liquid compared with
the same volume of water. When applying centrifugal pumps, pressures
should be expressed in units of feet (meters) of head rather than in psi (kPa).
As an example of the equivalency between psi and feet of head demonstrated by Equation 1.1, atmospheric pressure at sea level can be expressed as
14.7 psi, or as 34 ft of water.
For positive displacement (PD) pumps (discussed in Section VI to follow),
the conversion to feet of head is not made, and pressures are expressed in
psi (kPa).
IV. Classification of Pumps
There are many ways to classify pumps: according to their function, their
conditions of service, materials of construction, etc. The U.S. pump industry
trade association, the Hydraulic Institute, has classified pumps as shown in
Figure 1.3. This classification divides pumps as follows:
A. Principle of Energy Addition
The first classification is according to the principle by which energy is added
to the liquid. There are two broad classes of pumps, defined below.
1. Kinetic
In a kinetic (also called rotodynamic) pump, energy is continuously added to
the liquid to increase its velocity. When the liquid velocity is subsequently
reduced, this produces a pressure increase. Although there are several special types of pumps that fall into this classification, for the most part this
classification consists of centrifugal pumps.
2. Positive Displacement
In a PD pump, energy is periodically added to the liquid by the direct application of a force to one or more movable volumes of liquid. This causes an
increase in pressure up to the value required to move the liquid through
ports in the discharge line. The important points here are that the energy
addition is periodic (i.e., not continuous) and that there is a direct application
5
Introduction to Pumps
Air
operated
Reciprocating
pumps
Steam
Power
Controlled
volume
Diaphram
Bellows
Piston
Horizontal
Vertical
Simplex
Duplex
Double-acting
Double-acting
Plunger
Horizontal
Packed plunger
Simplex
Packed piston
Duplex
Vertical
Flexible
member
Lobe
Circumferential
piston
Diaphram
Single
Multiple
External
Internal
Single
Multiple
Single
Multiple
Impeller between
bearings
Turbine type
axial flow designs.
Regenerative
turbine
Special
effect
Multiplex
Manual
control
Automatic
control
Flexible tube
Flexible vane
Flexible liner
Centrifugala
aIncludes radial, mixed flow and
Mechanically
coupled
Hydraulically
coupled
Duplex
Multiplex
Blade, bucket
roller or
slipper
Axial
Radial
Overhung
impeller
Kinetic
Simplex
Vertical
Gear
Pumps
Duplex
Piston
Piston
Screw
Simplex
Plunger
Single-acting
Vane
Rotary
pumps
Piston
Horizontal
Blow case
Positive
displacement
Double-acting
Single-acting
Close-coupled
single and two stage
End suction
(including submersibles)
In-line
Separately coupled
single stage
In-line
Frame mounted
Centering support
API 610
Frame mounted
ANSI B73.1
Wet pit volute
Axial flow impeller (propeller)
volute type (horizontal or
vertical)
Axial (horizontal) split case
Radial (vertical) split case
Separately coupled
multistage
Axial (horizontal) split case
Radial (vertical) split case
Separately coupled
single and two stage
Vertical type
single and multistage
Overhung impeller
Impeller between bearings
Deep well turbine
(including submersibles)
Barrel or can pump
Short setting or close-coupled
Axial flow impeller (propeller)
or mixed flow type (horizontal
or vertical)
Single stage
Two stage
Reversible centrifugal
Rotating casing (pitot tube)
FIGURE 1.3
Classification of pumps. (Courtesy of Hydraulic Institute, Parsippany, NJ; www.pumps.org.)
6
Pump Characteristics and Applications
of force to the liquid. This is most easily visualized through the example of
a reciprocating piston or plunger pump (see Figure 1.23). As the piston or
plunger moves back and forth in the cylinder, it exerts a force directly on the
liquid, which causes an increase in the liquid pressure.
B. How Energy Addition Is Accomplished
The second level of pump classification has to do with the means by which
the energy addition is implemented. In the kinetic category, the most common arrangement is the centrifugal pump. Other arrangements include
regenerative turbines (also called peripheral pumps) and special pumps such as
jet pumps that employ an eductor to bring water out of a well.
In the PD category, the two most common subcategories are reciprocating
and rotary pumps.
C. Geometry Used
The remaining levels of pump classification shown in Figure 1.3 deal with
the specific geometry used. With centrifugal pumps, the geometry variations have to do with the support of the impeller (overhung impeller vs. impeller between bearings), rotor orientation, the number of impellers or stages, how
the pump is coupled to the motor, the pump bearing system, how the pump
casing is configured, and pump mounting arrangements.
With PD pumps, as is discussed in more detail in Section VI, there are
many different types of rotary and reciprocating pumps, each with a unique
geometry.
V. How Centrifugal Pumps Work
Stripped of all nonessential details, a centrifugal pump (Figure 1.4) consists
of an impeller attached to and rotating with the shaft and a casing that encloses
the impeller. In a centrifugal pump, liquid is forced into the inlet side of the
pump casing by atmospheric pressure or some upstream pressure. As the
impeller rotates, liquid moves toward the discharge side of the pump. This
creates a void or reduced pressure area at the impeller inlet. The pressure
at the pump casing inlet, which is higher than this reduced pressure at the
impeller inlet, forces additional liquid into the impeller to fill the void.
If the pipeline leading to the pump inlet contains a noncondensable gas
such as air, then the pressure reduction at the impeller inlet merely causes
the gas to expand, and suction pressure does not force liquid into the impeller inlet. Consequently, no pumping action can occur unless this noncondensable gas is first eliminated, a process known as priming the pump.
7
Introduction to Pumps
Point of entrance
to impeller vanes
Flow
line
Flow
line
Volute
Impeller
Typical pump section
Section through impeller and
volute along mean flow surface
FIGURE 1.4
Centrifugal pump with single volute casing. (From Karassik, I.J. et al., Pump Handbook, 4th Ed.,
McGraw-Hill, Inc., New York, 2008. With permission.)
With the exception of a particular type of centrifugal pump called a selfpriming centrifugal pump, centrifugal pumps are not inherently self-priming if
they are physically located higher than the level of the liquid to be pumped.
That is, the suction piping and inlet side of centrifugal pumps that are not
self-priming must be filled with noncompressible liquid and vented of
air and other noncondensable gases before the pump can be started. Selfpriming pumps are designed to first remove the air or other gas in the suction line, and to then pump in a conventional manner.
If vapors of the liquid being pumped are present on the suction side of
the pump, this results in cavitation, which can cause serious damage to the
pump. Discussed in greater detail in Chapter 2, Section VI, cavitation may
also cause the pump to lose prime.
Once it reaches the rotating impeller, the liquid entering the pump moves
along the impeller vanes, increasing in velocity as it progresses. The vanes in
a centrifugal pump are usually curved backward to the direction of rotation.
Some special types of pump impellers (Chapter 7, Section VII) have vanes
that are straight rather than curved. The degree of curvature of the vanes
and number of vanes, along with other factors, determines the shape and
characteristics of the pump performance curve, which is described in Chapter
2, Section IV.
When the liquid leaves the impeller vane outlet tip, it is at its maximum
velocity. Figure 1.5 illustrates typical velocity and pressure changes in a centrifugal pump as the liquid moves through the flow path of the pump. After
the liquid leaves the impeller tip, it enters the casing, where an expansion
of cross-sectional area occurs. This expanded area is offset in many pumps
by the additional flow being directed into the casing by the rotating impeller vanes, so that there is an area along the flow path where velocity and
8
Pump Characteristics and Applications
Pressure
(PSI)
Pressure
Outlet tip of
impeller vane
Inlet tip of
impeller vane
Velocity
(ft/sec)
Suction
Velocity
Flow path
Discharge
FIGURE 1.5
Velocity and pressure levels vary as the fluid moves along the flow path in a centrifugal pump.
pressure are neutral in this part of the casing. This is illustrated by the horizontal portions of the velocity and pressure lines in Figure 1.5.
Continuing on through the flow path in the remainder of the casing, the
casing design ensures that the cross-sectional area of the flow passages
increases as the liquid moves through the casing. Because the area is increasing as the liquid moves through the path of the casing, a diffusion process
occurs, causing the liquid’s velocity to decrease, as Figure 1.5 illustrates. By
the Bernoulli equation (see Ref. [1] at the end of this book), the decreased
kinetic energy is transformed into increased potential energy, causing the
pressure of the liquid to increase as the velocity decreases. The increase of
pressure while velocity is decreasing is also illustrated in Figure 1.5.
A centrifugal pump operating at a fixed speed and with a fixed impeller
diameter produces a differential pressure, or differential head. Head is usually
expressed in feet or meters, and abbreviated TH (total head). The amount of
head produced varies with the flow rate, or capacity delivered by the pump, as
illustrated by the characteristic head–capacity (H–Q) curve shown in Figure 1.6.
As the head developed by the pump decreases, the capacity increases.
Alternatively, as the pump head increases, the flow decreases. Pump capacity
is usually expressed in gallons per minute (gpm) or, for larger pumps, in cubic
feet per second (cfs). Metric equivalents, depending on the size of the pump,
are cubic meters per second, liters per second, or cubic meters per hour.
The centrifugal pump casing is one of several types. A single volute casing
is illustrated in Figure 1.4. Note that the single volute casing has a single
9
Introduction to Pumps
Total head or total differential pressure
Slip
Centrifugal
Positive displacement
Flow rate
FIGURE 1.6
Typical head–capacity relationship for centrifugal and positive displacement pumps.
cutwater where the flow is separated. As the flow leaves the impeller and
moves around the volute casing, the pressure increases. This increasing pressure as the liquid moves around the casing typically produces an increasing
radial force at each point on the periphery of the impeller, due to the pressure acting on the projected area of the impeller. Summing all of these radial
forces produces a net radial force that must be carried by the shaft and radial
bearing system in the pump. The radial bearing must also support the load
created by the weight of the shaft and impeller.
The radial bearing loads generated by a pump also vary as the pump operates at different points on the pump performance curve, with the minimum
radial force being developed at the best efficiency point (BEP) of the pump
(Figure 1.7). See Chapter 2, Section V, for a discussion of BEP. Operation at
points on the pump curve to the right or to the left of the BEP produces
higher radial loads than are produced when operating at the BEP. This is
especially true of single volute casing pumps, as Figure 1.7 illustrates.
Symptoms of excessive radial loads include excessive shaft deflection and
premature mechanical seal and bearing failure. Continuous operation of
the pump at too low a minimum flow is one of the most common causes
of this type of failure. Because for rolling element bearings, bearing life is
inversely proportional to the cube of the bearing load, operating well away
from the pump BEP can cause a reduction in bearing life by several orders
of magnitude.
A diffuser (shown in orange in Figure 1.8) is a more complex casing arrangement, consisting of multiple flow paths around the periphery of the impeller. The liquid that leaves the impeller vanes, rather than having to move
completely around the casing periphery as it does with the single volute
10
Pump Characteristics and Applications
Radial
load (#)
Single volute
Double volute
Diffuser
Best efficiency point
H
(ft)
Q (gpm)
FIGURE 1.7
Typical radial loads produced by single volute, double volute, and diffuser casings.
Casing
Impeller
Diffuser
FIGURE 1.8
Diffuser casing minimizes radial loads in a centrifugal pump.
casing, merely enters the nearest flow channel in the diffuser casing. The
diffuser casing has multiple cutwaters, evenly spaced around the impeller,
as opposed to the one cutwater found in a single volute casing. The main
advantage of the diffuser casing design is that this results in a near balancing
of radial forces, thus reducing shaft deflection and eliminating the need for
a heavy-duty radial bearing system. The dead weight of the rotating element
Introduction to Pumps
11
must still be carried by the radial bearing, but overall the diffuser design
minimizes radial bearing loads compared with other casing types.
Because the diffuser design produces minimal radial bearing loads, one
might wonder why all pumps do not have diffusers rather than volute type
casings. The reason is partially due to economics, as a pump with a diffuser
casing generally has more parts or more complex parts to manufacture than
a pump with a volute casing. Depending on pump size and materials of construction, economics often do not justify the use of diffuser casings except
where significant savings can be achieved in the size of shaft or radial bearing that is used in the pump. This is usually only found to be the case in
multistage, high-pressure pumps. However, with multistage pumps there
are other considerations as well. Volute designs in some multistage pumps
allow, by the use of a cross-over, some of the impellers to be oriented in
the opposite direction, providing balancing of axial thrust loads. (Refer to
Chapter 4, Section II.D, and Section VII.) The leading manufacturers of multistage pumps are themselves not in agreement on this subject.
Vertical turbine pumps (Chapter 4, Section XI) usually have diffuser casings.
Because the bearings for these vertical pumps are submerged in the liquid
being pumped, it is not practical to have a ball- or roller-type radial bearing
for this type of pump. Rather, the radial bearing loads must be accommodated by a wetted sleeve type bearing, which is not an ideal bearing system
in this type of arrangement. Therefore, to minimize radial bearing loads,
diffuser type casings are used in this type of pump.
A hybrid between a single volute casing and a diffuser casing is a double
volute casing (Figure 1.9). With this casing design, the volute is divided, which
creates a second cutwater, located 180° from the first cutwater. This design
results in much lower radial loads than are present with single volute designs
(Figure 1.7). Double volute casings are usually used by pump designers for
larger, higher flow pumps (usually for flows greater than about 1500 gpm) to
allow the use of smaller shafts and radial bearings.
FIGURE 1.9
Double volute casings are used in larger centrifugal pumps to reduce radial loads. (Courtesy of
Goulds Pumps, Inc., a subsidiary of ITT Corporation.)
12
Pump Characteristics and Applications
VI. PD Pumps
A. General
This book is primarily about centrifugal pumps. However, as Figure 1.3 illustrates, there is an entire other class of pumps known as Positive Displacement
(PD) pumps that deserves some attention. One of the earliest decisions that
must be made in designing a system and applying a pump is the selection of
the type of pump to be used. The first issue is the general decision whether
the pump should be of the centrifugal or the PD type. Surveys of equipment engineers and pump users indicate that the majority of them have a
strong preference for centrifugal pumps over PD pumps (if the hydraulic
conditions are such that either type can be considered). Many reasons are
given for this preference for centrifugals, but most are related to the belief
that centrifugal pumps are more reliable and result in lower maintenance
expense. Centrifugal pumps usually have fewer moving parts, have no
check valves associated with the pumps (as reciprocating PD pumps do),
produce minimal pressure pulsations, do not have rubbing contact with the
pump rotor, and are not subject to the fatigue loading of bearings and seals
that the periodic aspect of many PD pumps produce. Centrifugals should
be considered first when applying a pump, but they are not always suited to
the application.
B. When to Choose a PD Pump
This preference for centrifugal over PD pumps is certainly not always the
case, and in fact, there are certain application criteria that demand the use of
a PD pump. The following are some key application criteria that would lead
to the selection of a PD pump over a centrifugal pump:
• High viscosity
• Self-priming
• High pressure
• Low flow
• High efficiency
• Low velocity
• Low shear
• Fragile solids handling capability
• Sealless pumping
• Accurate, repeatable flow measurement
• Constant flow/variable system pressure
• Two-phase flow
13
Introduction to Pumps
The ability to pump viscous liquids is one of the most import attributes
of PD pumps. It is possible to handle low-viscosity liquids with centrifugal
pumps. However, pump efficiency degrades rapidly as viscosity increases
and there is an upper limit of viscosity above which it becomes impractical
to consider centrifugal pumps due to the excessive waste of energy. Highly
viscous liquids absolutely cannot be pumped with a centrifugal pump. (See
Chapter 3, Section II, for more discussion of viscosity and its effect on centrifugal pump performance.) For these liquids, some type of PD pump may
be the only practical solution.
Most PD pump types are inherently self-priming, meaning they can be
located above the surface of the liquid being pumped without the necessity
of the suction line being filled with liquid and the noncondensable gases in
the suction line being removed before starting the pump. Therefore, these
pump types can be conveniently mounted on top of transfer tanks with no
special external priming devices. (Refer to further discussion of the dry selfpriming capability of PD pumps in Section VI.C and Table 1.1 to follow.)
The high pressure and low flow criteria above must be considered together.
How high is high pressure, and how low is low flow? It is possible to find,
for instance, centrifugal pumps that produce pressures of several thousand
psi; and certainly, one can find very small centrifugal pumps whose capacity
is only a couple of gallons per minute. However, what if one has an application for 5 gpm at 2000 psi? In that case, a PD pump is about the only solution.
Figure 1.10 shows in very broad terms the head and flow range of centrifugal,
rotary, and reciprocating pumps. Although there is a portion of this coverage
100,000+
Reciprocating
H
(psi)
6500
Centrifugal
4500
Rotary
1800
15,000
Q (gpm)
FIGURE 1.10
Head vs. flow for centrifugal, rotary, and reciprocating pumps.
200,000+
= Scale change
14
Pump Characteristics and Applications
chart that can be met by all three pump types, the one area that stands out
as being able to be met only by a PD pump is low flows in combination with
very high pressures.
If energy efficiency were the only consideration in selecting pumps, more
PD pumps would be considered, since some PD pumps are quite energy efficient. Energy is not the only consideration though, and other factors such as
installed cost and maintenance expense often outweigh the energy savings.
The criteria of low velocity, low shear, and fragile solids handling capability often go hand in hand. Centrifugal pumps, because of the high velocities
present at the impeller discharge, and because of the close clearances inside
the pump, often subject the pumped liquid to high shear stresses. Many liquids cannot tolerate these high velocities and high shear stresses. A good
example of this is fruits and vegetables such as cherries and peas that are
pumped in food-processing plants. If these products were pumped using
centrifugal pumps, they would produce cherry juice and pea juice. There is
an entire class of centrifugal pump impellers of the nonclog type, whose function is to pump sewage and other waste liquids containing solids (Chapter
4, Section II.F). These centrifugal pumps, however, are not concerned with
maintaining the integrity of the solids and often shred the solids as they
pump them. Another centrifugal impeller type, known as a recessed, or vortex impeller (Chapter 4, Section II.F), is capable of pumping large solids with
minimal degradation, but the downside is that this impeller type is very,
very inefficient.
Because of their unique design, several types of PD pumps, such as peristaltic (Section VI.C.4) and diaphragm (Section VI.C.13) pumps, are inherently
sealless, requiring no shaft seals and having zero product leakage. Although
there are available today several types of sealless centrifugal pumps as
Chapter 5, Section V, discusses, these centrifugal options also have their limitations and shortcomings. The inherent sealless nature of some PD pumps
may make them a simpler solution.
With PD pumps, capacity varies directly with speed and is independent
of differential pressure or head. Figure 1.6 illustrates that pump capacity is
independent of differential pressure for PD pumps while being dependent
on differential pressure (head) for centrifugal pumps.
Most PD pumps exhibit slip, that is, leakage from the high pressure to the
low-pressure side of the pump. As shown in Figure 1.6, slip causes the pump
to deliver a lower flow rate at higher differential pressures. The amount of
slip varies widely from one PD pump to another, as well as varying with
pump differential pressure, with the liquid viscosity, and with internal clearances. Most PD pumps are not nearly as subject to increased leakage back to
suction because of wear as are centrifugal pumps. Some types exhibit very
little slip. These factors make some types of PD pumps ideal for metering
applications, where an accurate, controllable flow rate of (usually) an expensive chemical must be dispensed, for example, to treat water with chlorine.
Note that centrifugal pumps could also be used to accurately control flow
Introduction to Pumps
15
but they would have to rely on a control loop that measures flow and then
adjusts a system control valve. Reciprocating pumps are the most common
type of PD pump used for metering pumps, although other types such as
peristaltic pumps are also used for metering.
The requirement for a constant process flow rate where system pressure
varies widely can be met with a centrifugal pump. However, this usually
requires a feedback control system, including flow measurement and the use
of an automatic control valve to maintain constant flow. A PD pump may be a
much simpler solution because once the liquid characteristics and the speed
and size of a PD pump are fixed, the pump delivers a nearly constant flow
rate regardless of the system pressure against which the pump is pumping.
Finally, many PD pump types are much more suited to pumping liquids
containing gases than are centrifugal pumps.
One of the hydraulic characteristics of PD pumps is that the pump continues to deliver a constant flow rate (if pumping at a constant speed) while
building up pressure at the discharge (Figure 1.6). This means that if a valve
downstream of the pump is inadvertently closed, the pump continues to
build up pressure until something gives. Usually, one of several things happens to prevent damage to the pump. At constant flow and ever-increasing
pressure, the required horsepower may continue to build until the motor is
overloaded and trips off, or a high-pressure or high-temperature limit switch
may trip the motor. If the motor does not trip or the excessive pressure is not
relieved, the pump could eventually build up pressure until it overpressurizes the pump casing or some component downstream of the pump, causing potentially serious damage. Usually, it is recommended that PD pumps
incorporate a pressure relief valve, built into the pump or supplied just
downstream of the pump, to protect against overpressurization.
Internal relief valves built into the pump relieve pressure via an internal
loop that connects discharge to suction. External valves must be piped to an
external source, generally the supply reservoir. The external approach can
provide a visual indicating that the valve is open.
Relief valve settings are sometimes adjustable and can be tricky. The user
should be certain to understand the manufacturer’s definition of the valve
setting and the valve characteristics. These valves will wear very quickly
and generally are not stable if used as a pressure-regulating device. It is best
to obtain a pressure-regulating valve for this purpose.
C. Major Types of PD Pumps
The following pages provide a brief description of the most common types
of PD pumps. The first five types (sliding vane, sinusoidal rotor, flexible impeller,
flexible tube, and progressing cavity) are single-rotor rotary pumps. The next
five types (external gear, internal gear, rotary lobe, circumferential piston, and multiple screw) are multiple-rotor rotary pumps. The last three (piston, plunger,
and diaphragm) are reciprocating pumps.
16
Pump Characteristics and Applications
Table 1.1 summarizes the primary application criteria for PD pumps. In
using this table, the following guidelines should be observed.
• The information given is approximate only and is meant to provide
assistance in selecting appropriate pump types to consider for a particular application. The parameters of performance for any single
pump type vary widely from manufacturer to manufacturer and
from one pump design to another. In general, Table 1.1 shows the
limits of performance for the major pump types. In many cases,
there are only one or two manufacturers that make pumps with the
indicated maximum value of flow, pressure, or other performance
parameter. For any PD pump application, the best advice is to obtain
literature and performance data from several manufacturers of the
types of pump being considered.
• The parameters given for a particular pump type do not necessarily
apply concurrently. As examples of this, the maximum flow listed is
likely not available with the maximum pressure listed for that pump
type. The maximum solids size listed is probably only able to be
handled by the largest pumps of that type, and may be restricted
below the maximum size shown in the table for solids of a particular
shape, material, or concentration.
• Virtually all PD pumps are self-priming to a certain degree. Only
the few pump types so indicated in Table 1.1 are truly dry self-priming,
that is, will prime on a lift when completely dry.
• The columns of Table 1.1 that rate the various pump types in categories such as ability to handle abrasives and fragile solids have been
prepared using the author’s best judgment. The ratings may change
for particular applications, and arguments could easily be made to
move any pump up or down in the relative rankings.
• Some pumps are used for sanitary food-grade applications (manufactured in stainless steel, have food-grade elastomers, etc.), but do
not have USDA or 3A certification. Pumps without such certification
would be used in the upstream end of the food processing system,
prior to the sterilization process, or would otherwise be used in food
applications where such certification is not required.
• No attempt has been made in Table 1.1 to present any rating of the
pump types relative to cost. The reasons for this include the fact
that features and relative costs are so application specific; that many
of these pump types were never intended to compete against each
other; and the fact that any good cost comparison must include the
costs associated with installation, space requirements, maintenance
and repair parts, and energy.
Sliding vane
Sinusoidal rotor
Flexible impeller
Flexible tube (peristaltic)
Progressing cavity
External gear
Internal gear
Rotary lobe
Circumferential piston
Two-screw
Three-screw
Piston
Plunger
Diaphragm
Air-operated diaphragm
Wobble plate
Pump Type
2500
300
150
200
2400
1200
1500
3000
600
15,000
4500
700
1200
1800
300
50
Maximum
Capacity (gpm)
200
200
60
220
2000
2500
200
450
200
1500
4500
5000
100,000
17,500
125
1500
Maximum
Pressure (psi)
Key Application Data of Positive Displacement Pumpsa
TABLE 1.1
0.5
18.0
0.1
0.2
5.00
2.00
2.00
5.00
5.00
4.05
1.00
0.05
0.05
1.00
0.75
0.025
Maximum
Viscosity
(million SSU)
0.031
2
1
1
2
—b
—b
4
2
—b
—b
0.50
0.50
1
2
0.125
Maximum
Solid Size (in)
N
N
Y
Y
Y
N
N
N
N
N
N
Y
Y
Y
Y
Y
Dry SelfPriming (Y/N)
28
30
24
30
30
20
20
20
20
31
28
25
20
14
25
8
(continued)
Maximum Suction
Lift (ft H2O)
Introduction to Pumps
17
c
b
a
Y
N
N
Y
N
N
N
Y
Y
Y
N
N
N
Y
Y
Y
3
4
2
1
1
5
5
2
2
3
4
2
4
1
1
1
Abrasive
Handling
Ratingc
Refer to guidelines in text for the use of this table.
Friable solids only or rotors must be hardened and clearances opened.
Rating of 1 is best, 3 is medium, 5 is worst.
Sliding vane
Sinusoidal rotor
Flexible impeller
Flexible tube (peristaltic)
Progressing cavity
External gear
Internal gear
Rotary lobe
Circumferential piston
Two-screw
Three-screw
Piston
Plunger
Diaphragm
Air-operated diaphragm
Wobble plate
Pump Type
Able to Run Dry
(Y/N)
Key Application Data of Positive Displacement Pumpsa
TABLE 1.1 (Continued)
3
1
2
1
1
4
4
1
1
4
5
3
3
2
2
3
Fragile Solids/
Shear Sensitive
Liquidsc
3
1
2
4
1
1
1
3
3
1
1
5
5
5
5
3
Pulsationsc
3
3
5
2
2
3
3
3
3
4
4
1
1
1
5
1
Metering
Abilityc
N
Y
Y
Y
Y
N
Y
Y
Y
N
N
N
Y
Y
Y
N
Sanitary Designs
Available (Y/N)
18
Pump Characteristics and Applications
Introduction to Pumps
19
For the reader who is interested in finding further information about a
specific type of pump, Appendix A at the end of this book contains a listing
of most of the major pump suppliers in the United States (both domestically
produced and imported into the U.S.) segmented according to the pump
types they offer. This list is by no means all-inclusive, containing only those
manufacturers with which the author is familiar. Readers interested in locating information on a particular pump type should be able to use Appendix
A as a guide, and should be able to locate manufacturers via an Internet
search, in a Thomas Register, or similar index of manufacturers. Note that
Appendix A lists the manufacturers alphabetically, and that it includes both
centrifugal and PD pump suppliers.
The major types of PD pumps are described below. For each type of pump,
the primary application characteristics are noted, as well as the benefits and
shortcomings of each type.
1. Sliding Vane Pump
In a sliding vane style of rotary pump, shown in Figure 1.11, vanes cooperate with a cam to draw liquid into and force it from the pump chamber.
In this pump style, vanes fit into slots cut lengthwise in the rotor, and the
rotor turns inside an eccentrically shaped casing that acts like a cam. When
the rotor is turning at operating speed, the vanes are forced by centrifugal
force outward until they come in contact with the casing wall. Some types of
vane pumps also rely on springs to force the vanes outward, so that contact
between the vanes and the casing walls is maintained even when the pump
is operating at slow speeds. Other types use liquid pressure from the pump
discharge acting beneath the vanes to force the vanes outward.
Advantages of vane pumps include their rather simple construction, the
fact that they are self-compensating for wear on the vanes, and that they
operate well with thinner, low-viscosity liquids. They can operate with
mildly erosive liquids. The disadvantages of this pump type include their
FIGURE 1.11
Sliding vane pump. (Courtesy of Blackmer, a Dover Company, Grand Rapids, MI.)
20
Pump Characteristics and Applications
inability to pump highly viscous liquids and the fact that they cannot handle
fragile solids.
2. Sinusoidal Rotor Pump
Figure 1.12 shows a sinusoidal rotor pump. In this pump type, a rotor having the shape of two complete sine curves turns in a housing, creating four
separate, symmetrical pumping compartments. A sliding scraper gate covers
part of the rotor, oscillates as the rotor turns, and prevents return of product
past the discharge and back to the suction side of the pump.
The major benefits of the sinusoidal rotor pump are its low shear and its
gentle handling of fragile solids and highly viscous liquids. Its principal
shortcoming is that it has limited ability to handle highly abrasive liquids.
3. Flexible Impeller Pump
In the flexible impeller pump (Figure 1.13), sometimes called a flexible vane
pump, the rotor is made of an elastomeric material such as rubber. The
blades of this “impeller” continuously deflect and straighten as they pass
across a cam between the inlet and discharge ports. The flexing of the blades
produces a vacuum that causes liquid to flow into the space between the two
blades and then moves the liquid through the pump.
Advantages of the flexible impeller pump include the fact that it is dry
self-priming; can handle liquids with solids, abrasives, or entrained air; and
is relatively inexpensive. Disadvantages include the upper limits of flow
FIGURE 1.12
Sinusoidal rotor pump. (Courtesy of Watson-Marlow Pumps Group, Wilmington, MA.)
21
Introduction to Pumps
1
2
3
FIGURE 1.13
Flexible impeller pump. (Courtesy of Jabsco, a Xylem brand, Beverly, MA.)
(about 150 gpm) and pressure (about 60 psi) and the fact that the pump cannot run dry longer than a couple of minutes without doing damage to the
rubber impeller.
4. Flexible Tube (Peristaltic) Pump
The flexible tube pump (Figure 1.14) is also called a peristaltic pump, or simply a hose pump. In this pump type, a flexible tube made of rubber or other
material is located inside a circular housing. Rollers or cams attached to the
rotor squeeze the tube as they pass across it, drawing the liquid through the
pump. This action is similar to what happens when a person swallows, a
process called peristalsis, which is how the pump gets its name.
Peristaltic pump advantages include the fact that they are sealless (require
no packing assembly or mechanical seal), can handle quite corrosive liquids
(as long as the tube material is compatible with the liquid being pumped),
are dry self-priming, and are relatively inexpensive. They are frequently
used as low-cost metering pumps, for applications such as chlorine metering
in commercial swimming pools.
FIGURE 1.14
Peristaltic (hose) pump. (Courtesy of Watson-Marlow Pumps Group, Wilmington, MA.)
22
Pump Characteristics and Applications
Disadvantages include relatively low flow and pressure capability for most
models (although several manufacturers offer pressures to several hundred
psi), and hoses usually require changing about every three months. The
selection of the proper hose material for the application is the most critical
aspect of applying the hose pump. Finally, peristaltic pumps are not as accurate for metering as reciprocating style pumps.
5. Progressing Cavity Pump
The progressing cavity (PC) pump in its most common design has a singlethreaded screw or rotor, turning inside a double-threaded stator (Figure 1.15).
The stator is made of an elastomeric material, and the rotor has an interference fit inside the stator. As the rotor rotates inside the stator, cavities form at
the suction end of the stator, with one cavity closing as the other opens. The
cavities progress axially from one end of the stator to the other as the rotor
turns, moving the liquid through the pump.
Advantages of PC pumps include their ability to pump highly viscous
liquids (as well as very low viscosity liquids), shear sensitive liquids, fragile
solids, and abrasives. Also, the pump produces very little pulsation and is selfpriming even when dry. The maximum pressure is limited to about 75 psi, but
the PC can be set up with many stages in series using the same driver, so it can
achieve upper pressure limits well over 1000 psi. Another advantage is that the
packing or seal sees suction pressure rather than discharge pressure.
Disadvantages of the PC pump include the relatively higher cost of replacement parts, its large floor space requirement, the fact that the pump cannot
run dry for an extended period, and upper temperature limits of about 300°F.
(The material of the elastomeric stator usually sets the upper temperature
limit.) Starting torques are quite high, so the pump motor may need to be
much larger than would be the case with other types of PD pumps.
Note that as the rotor turns, the centerline of the rotor orbits about the
centerline of the stator. There are several methods to connect the PC pump’s
drive shaft to its rotor and to account for the elliptical motion of the rotor
shaft. Two common configurations are the pin drive connecting rod and
the crowned gear drive connecting rod (the latter is shown in Figure 1.15).
FIGURE 1.15
Progressing cavity pump. (Courtesy of Moyno, Inc., Springfield, OH.)
Introduction to Pumps
23
Although both configurations are universally accepted, they differ in significant ways. The pin drive configuration is the most common and least
expensive of the two types. The crowned gear drive configuration is not as
common and is more expensive than the pin drive; however, it is significantly more heavy duty. The crowning of the gears permits this coupling to
accept the elliptical motion of the rotor. The crowned gears also provide a
large surface area to transfer the torque load required to turn the rotor.
Variations on progressing cavity designs use more threads on the rotor
and stator or special stator materials.
6. External Gear Pump
External gear pumps (Figure 1.16) have two meshing gears, which may be
of the spur, helical, or herringbone type. These three gear types are illustrated
in Figure 1.17. Liquid is carried between the gear teeth and displaced as the
teeth mesh. Close clearances between the gear teeth and between the teeth
and the casing walls minimize slippage of liquid from the high-pressure side
to the low-pressure side.
Spur gears are simple, and generally cost less to manufacture than helical or herringbone gears. They have good characteristics but can be noisy
and inefficient. The trapped liquid between the teeth where the teeth mesh
has no place to exit. As the trapped liquid squeezes by the tight clearances,
it can make a loud screaming sound. Some manufactures provide a relief
slot in this area to give the fluid a place to escape. This minimizes noise
and increases efficiency, especially with viscous liquids. Spur gears have the
added advantage of minimal axial thrust.
Helical gears (e.g., Figure 1.16) are another means of giving the trapped liquid a path to escape. The helix shape gives the liquid a place to exit. Helical
gear pumps are generally efficient and quiet. They do have one disadvantage: they are forced axially against the pump thrust bushing. Axial wear
FIGURE 1.16
External gear pump. (Courtesy of Roper Pump Company, Commerce, GA.)
24
Pump Characteristics and Applications
β
Spur
gear
Helical
gear
Herringbone
gear
FIGURE 1.17
Spur, helical, and herringbone gears. (Courtesy of Diener Precision Pumps, Lodi, CA.)
in a gear pump decreases performance much faster than radial wear, so it is
important to maintain tight clearances in the axial direction.
Herringbone gears are the most expensive to manufacture, but they are quiet,
efficient, and do not exhibit axial thrust. They are very difficult to machine,
so sometimes they are made up with two helical gears butted together with
the helix of the two gears oriented opposite to each other.
Most external gear pumps have the first gear (which is coupled to the driver)
driving the second gear. An alternative design uses timing gears, which are a
separate set of gears, normally isolated from the pumped liquid, which cause
the two main gears to remain in mesh without requiring direct contact. This
use of timing gears, common on many types of double-rotor rotary pumps,
allows larger clearances between the pumping gears, which means that the
pump can more easily accommodate some abrasives without doing damage
to the gears. Timing gears are shown on the pump in Figure 1.20.
Advantages of external gear pumps include the fact that they operate at
relatively high speeds, producing relatively high pressures. The gears are
usually supported by bearings on both sides, so there are no overhung loads.
The pumps are relatively quiet (particularly if helical or herringbone gears
are used), largely free of pulsations, and economical. Disadvantages include
the fact that there are usually four bushings in the liquid area, and, unless
timing gears are used, these pumps cannot tolerate solids or abrasives.
The weakest element in most gear pumps (external gear or internal gear
type) is the driven gear bearing. One might think that the driving gear
Introduction to Pumps
25
carries 100% of the load, so the driving bearings should wear the most, but
this is not the case.
Radial bearing loads are a result of pump differential pressure and the
forces of the driving gear pushing the driven gear. This produces higher
loading on the driven gear bearing than on the driving gear bearing. This
problem is usually exacerbated by the fact that the driven gear bearing is
wetted by the fluid being pumped.
So when examining a gear pump for wear, look carefully at the driven gear
bearings first. As they wear, the gear is no longer supported and the teeth
will begin to wear against the pump wall or wear plate. The mesh between
gears will increase, causing slip and reduced pump performance, especially
at high pressures.
Considering the materials of construction of gears, it should be considered
that most gear pumps pump lubricating fluids. For these pumps, the gears
are made of a variety of metals, but mostly steel. Hardness runs from that of
mild steel to very hard treated steels to produce high pressures.
For nonlubricating fluids, bronze and a variety of nonmetallic materials such as plastic composites, carbon, Teflon®, and ceramics are used. The
choice is often dictated by chemical resistance, mechanical properties, and
cost. For example, Teflon has excellent chemical resistance but its physical
properties often result in excessive wear, low-pressure capability, and fluid
temperature is limited (high thermal expansion).
Due to the tight clearances required in gear pumps, there is no allowance
for abrasive wear. Most gear pumps will not handle abrasives without premature wear failure. Some, however, are made of wear-resistant materials
and can perform quite well for some applications. These pumps use hardened shafts, ceramic gears and/or bushings, or highly abrasion-resistant
plastic composites. Running the pump very slowly is another means of getting more life out of a pump under all conditions but is purposely done for
abrasive applications.
7. Internal Gear Pump
Internal gear pumps (Figure 1.18), like external gear pumps, move and pressurize liquid by the meshing and unmeshing of gear teeth. With an internal gear pump, a rotor having internally cut teeth meshes with and drives
an idler gear having externally cut teeth. Pumps of this type usually have
a crescent-shaped partition that moves the liquid through the pump with
minimal slip.
Internal gear pumps’ advantages include their few moving parts, relatively low cost, and the fact that they have only one seal. They can usually
operate well in either direction, and reversing rotation causes a reversal in
the direction of flow. Disadvantages include the fact that there is one bearing
in the pumped liquid, that the one bearing must support an overhung load,
and these pumps generally will not work with abrasives or solids.
26
Pump Characteristics and Applications
FIGURE 1.18
Internal gear pump. (Courtesy of Viking Pump, Inc., a unit of Idex Corporation.)
8. Rotary Lobe Pump
A lobe pump (Figure 1.19) is similar to an external gear pump, in that the liquid is carried between the rotor lobe surfaces that cooperate with each other
as they rotate to provide continuous sealing, as do the teeth of a gear pump.
Unlike a gear pump, however, one lobe cannot drive the other, so this type
of pump must have timing gears (described in Section VI.C.6 and shown in
Figure 1.20) to allow the lobes to remain in synch with each other. Each rotor
may have one lobe or several, with three lobes being most common.
FIGURE 1.19
Lobe pump. (Courtesy of Viking Pump, Inc., a unit of Idex Corporation.)
Introduction to Pumps
27
FIGURE 1.20
Bi-wing lobe pump. (Courtesy of Viking Pump, Inc., a unit of Idex Corporation.)
The wide spaces between the lobes and the slow speeds at which these
pumps operate make this style of pump ideal for handling food products
containing fragile solids. Shear is low with this pump. Also, the fact that the
pump has timing gears means that the lobes do not actually touch each other
as they rotate, a necessary condition for many sanitary food-grade applications. Disadvantages include the fact that these pumps are subject to pressure pulsations, and they have fairly high amounts of slip with low-viscosity
liquids. Finally, they must have timing gears.
Note that some manufacturers make these pumps with elastomeric lobes,
whereas others make them with metal lobes. The different lobe materials
affect maximum pressure and temperature, as well as the price of the pump.
9. Circumferential Piston, Bi-Wing Lobe Pumps
The circumferential piston and bi-wing lobe pumps are very similar to the
traditional lobe pump, both in the way they operate and in their applications. Figure 1.20 shows a bi-wing lobe pump. Instead of traditional lobes,
the rotors have arc shaped “pistons,” or rotor wings, traveling in annularshaped “cylinders” machined in the pump body. As with traditional lobe
pumps, the rotors are not in direct contact with each other and require the
use of timing gears.
Circumferential piston and bi-wing lobe pumps have less slip than do
comparably sized traditional lobe pumps. This is because the rounded lobes
of a traditional lobe pump only come in close contact with the casing at a
28
Pump Characteristics and Applications
single point on the outer surface of each lobe, whereas the circumferential
piston and bi-wing lobe pumps have a close clearance between the rotor and
the casing over the entire length of the arc. The lower slippage means that
these pumps are more energy efficient than traditional lobe pumps. With
liquid viscosities greater than about 2000 SSU (Chapter 3, Section II), this
advantage disappears.
Advantages and disadvantages of circumferential piston and bi-wing lobe
pumps are pretty much the same as for traditional lobe pumps, and again,
food processing is the most common application. Circumferential piston and
bi-wing lobe pumps normally do not handle abrasives as well as traditional
lobe pumps, and the arc-shaped rotor may make contact with the casing at
times in higher pressure applications.
10. Multiple-Screw Pump
This pump is called a multiple-screw pump to distinguish it from a single-­
screw (progressing cavity) pump. In a multiple-screw pump, liquid is carried
between rotor screw threads and is displaced axially as the screw threads
mesh. Multiple-screw pumps can be either two-screw (Figure 1.21) or threescrew (Figure 1.22). Two-screw pumps usually have timing gears, and
because the screws are not in contact with each other, this style can handle
more abrasive liquids than three-screw pumps. Three-screw pumps are usually untimed, with the central screw driving the other two screws. Because
the screws are in contact with each other, this style of pump cannot tolerate
severe abrasives, though these pumps are used to pump crude oil containing
sand. The screws on a two-screw pump are opposed, so that they balance out
all axial loads. Three-screw pumps, on the other hand, generate some axial
loads, although this is usually partially or fully balanced out.
FIGURE 1.21
Two-screw Warren Pump. (Courtesy of Colfax Fluid Handling, Monroe, NC.)
Introduction to Pumps
29
FIGURE 1.22
Three-screw Imo Pump. (Courtesy of Colfax Fluid Handling, Monroe, NC.)
The three-screw pump is usually the most economical choice if the liquid is not overly abrasive, because it has fewer parts (no timing gears). Both
types of screw pump have an extremely high upper limit of viscosity, and
screw pumps in general produce higher flow rates than any other type of
rotary PD pump. Screw pumps operate at high speeds and can produce very
high pressures with almost no pulsations and with relatively quiet operation. Two-screw pumps are usually limited to about 1500 psi, whereas threescrew pumps can go up to about 4500 psi.
11. Piston Pump
Piston pumps are reciprocating pumps consisting of a power end and a liquid
end. The power end of the pump takes the driver power and converts the rotary
motion to reciprocating motion. The driver can be a motor, an engine, or a turbine, whose output must then be converted by means of crankshafts, connecting rods, and crossheads, to reciprocating motion. An alternative, older
design called direct-acting has a steam cylinder for the power end, with inlet
and outlet valves to allow steam to directly drive the liquid end piston.
The liquid end of a reciprocating piston pump consists of a chamber having liquid inlet and outlet ports, with most designs having check valves in
both the inlet and outlet ports. When the reciprocating piston strokes in one
direction, the inlet check valve opens (while the outlet check valve remains
closed), directing liquid into the liquid end of the pump. When the piston
strokes in the opposite direction, the inlet check valve closes and the outlet
check valve opens, allowing liquid to move into the discharge port and out
into the system.
Piston pumps have the pistons operating inside cylinders, with the sealing taking place by means of elastomeric packing or O-rings on the outside
of the piston. Alternatively, the piston itself may be made of an elastomer
with an interference fit. Piston pumps can be single-acting or double-acting,
although most are double-acting. With a double-acting piston pump, liquid
30
Pump Characteristics and Applications
is discharged during both the forward and return motions of the piston
(requiring four check valves for each piston). A single-acting piston pump
has liquid discharging only during the forward stroke of the piston. Doubleacting designs are generally associated with slower speeds and medium
pressures, while single-acting designs are more generally associated with
higher speeds and higher pressures.
Note that high inlet pressures are a very important consideration in the
selection of a piston pump. Double-acting piston pumps are quite a bit more
efficient than single-acting pumps when the inlet pressures are quite high.
Double-acting piston pumps directly utilize the inlet pressure to reduce the
power end load. Also, high inlet pressures can affect the selection of a prime
mover. Double-acting piston pumps can use the pressure differential in the
calculation of the horsepower requirement, whereas single-acting pumps
can only use a fraction of the inlet pressure to offset discharge pressure.
Reciprocating pumps can have a single reciprocating piston, plunger, or
diaphragm (called simplex construction), or there may be multiple reciprocating components. A pump having two reciprocating members is called duplex
construction; one having three is called triplex; etc.
Piston pumps are used in applications where their high-pressure capability
makes them one of the only alternatives to consider. Relatively high-pressure
service with an abrasive liquid is one very common application. Most centrifugal pumps capable of handling abrasives have limited head capability
(only several hundred feet of head). Reciprocating piston pumps are considered better in abrasive applications than plunger pumps. With piston pumps
in abrasive applications, elastomeric pistons are used and cylinders are lined
with tungsten carbide or other hardened material.
Important applications are found for both piston and plunger pumps in
oil production and pipeline services. They are usually built in a horizontal
configuration for smaller sizes, and in vertical configurations for the larger
sizes to reduce the loads on the bearings, packing, and crossheads. Vertical
designs save floor space, but are relatively more expensive and more difficult
to maintain than horizontal designs.
Piston pumps are limited to slower speeds than plunger pumps (piston
pumps usually run at 100 rpm or slower) for several reasons. This is partly an
effort to reduce the high unbalanced forces from the two throws at 90°. Also,
piston pumps are usually designed for longer life than plunger pumps, so the
slower speed makes for a more conservative design. The slower speed of a
piston pump makes it usually larger in size and thus more expensive from a
capital standpoint, than a plunger pump for a particular rating. However, its
slower speed usually makes the piston pump a lower maintenance expense
alternative compared with a plunger pump.
Note that the allowable or rated speed of a reciprocating pump is not based
solely upon whether the pump is a double-acting piston pump or a singleacting piston or a plunger. The size of the valves limits the speed of the pump
for any given piston or plunger size. Sometimes, the size of the inlet opening
31
Introduction to Pumps
is also a factor in limiting the speed of the pump. Furthermore, the viscosity of the liquid also affects the speed of the pump, as reciprocating pumps
must be slowed down for more viscous liquids. The flow of liquid through
any check valve in a reciprocating pump is never steady but is constantly
changing from zero to a maximum. Therefore, considerable attention must
be given to the selection of a pump that does the job required without cavitation on the inlet stroke. (See Chapter 2, Section VI, for a discussion of NPSH
and cavitation.)
12. Plunger Pump
A plunger pump (Figure 1.23) is similar to a piston pump, except the reciprocating member is a plunger rather than a piston. The plunger is single-acting
(i.e., only has liquid discharging during the forward stroke), and the plunger
is sealed with packing in the cylinder walls.
Generally speaking, plunger pumps are used for higher pressure applications than piston pumps. They are capable of the highest pressures obtainable
with a PD pump, with some very special applications achieving pressures
greater than 50,000 psi. Plunger pumps generally run at higher speeds than
piston pumps. They are therefore usually a lower capital cost alternative to
a piston pump if both are being considered. However, the plunger pump
Check valve
(discharge)
Flushing
port
Pump head
Plunger holder cover
sealing
Plunger
Flushing port
Pump head holder
Check valve
(suction)
FIGURE 1.23
Plunger pump. (Courtesy of Lewa, Inc., Holliston, MA.)
Plunger rod
32
Pump Characteristics and Applications
may have higher maintenance expense and lower abrasion resistance than a
piston alternative.
For all types of reciprocating pumps, rather large pressure pulsations
are produced. These pulsations become more smoothed out the higher the
number of reciprocating members. Even so, quite often it is necessary to fit
pulsation dampener devices downstream of the pump discharge. A pulsation dampener is a vessel that is separated in the middle with a bladder or
membrane and that has air or nitrogen or a neutral gas on the upper half of
the vessel. The membrane flexes, and the air compresses and dampens out
the pulsations produced as the reciprocating pump strokes.
13. Diaphragm Pump
Diaphragm pumps (Figures 1.24–1.27) are similar to piston and plunger
pumps, except that the reciprocating motion of the pump causes a diaphragm to flex back and forth, which in turn causes the liquid to flow into
and out of the liquid end of the pump. As with all reciprocating pumps,
diaphragm pumps require check valves at the inlet and outlet ports (shown
as ball checks in Figures 1.24 and 1.26). The diaphragm is usually made of an
elastomeric material to allow it to flex. The diaphragm can be mechanically
attached to the reciprocating member or it can be separated and actuated by
a reservoir of hydraulic fluid, often with a contour plate to control the travel
limits of the diaphragm.
One very common application for diaphragm pumps of the type described
above is for metering applications. Metering pumps, or dosing pumps as
they are called in Europe, have relatively low flow rates, usually measured
in gallons or liters per hour rather than per minute. These pump types
are highly accurate in measuring flow (usually having an accuracy of better than ±1%), and the diaphragm makes the pump leak-free and compatible with a variety of liquids. Figure 1.24 shows a hydraulically actuated
diaphragm-metering pump with stroke adjustment capability to vary the
flow rate.
Another style of diaphragm pump is the solenoid metering pump,
used in light-duty metering applications. This style of diaphragm pump
uses an electrical signal to magnetically move the plunger/diaphragm
assembly.
Much larger versions of hydraulically actuated diaphragm pumps are
used in process services, where their high pressure capability and sealless
pumping make them an interesting alternative for special services. These
pumps, with metal diaphragms and remote heads, can pump liquids to
900°F.
Another type of diaphragm pump is the air-operated, double-diaphragm
pump (Figures 1.25 and 1.26). In this pump design, compressed air enters
the air chamber behind one of the diaphragms (shown in Figure 1.26), flexing the diaphragm and thus forcing the air or liquid on the other side of
Introduction to Pumps
FIGURE 1.24
Diaphragm metering pump. (Courtesy of Lewa, Inc., Holliston, MA.)
FIGURE 1.25
Air-operated, double-diaphragm pump. (Courtesy of Lutz Pumps, Norcross, GA.)
33
34
Pump Characteristics and Applications
(a)
(b)
FIGURE 1.26
Air-operated, double-diaphragm pump. (Courtesy of Warren Rupp, Inc., a unit of Idex
Corporation.)
the diaphragm out the discharge check valve. Simultaneously, the second
diaphragm (Figure 1.26) is pulled inward by a rod connecting the two diaphragms, creating a suction stroke with that diaphragm, with liquid coming
in through the inlet check valve on that side of the pump. Then a shuttle
valve causes the air distribution to shift, sending air to the other chambers
and reversing the stroke of the two diaphragms.
Any type of diaphragm pump has the distinct advantage, compared with
piston or plunger pumps, of being sealless, that is, not requiring any packing assembly or mechanical seal. Air-operated diaphragm pumps offer the
additional advantage of being able to accommodate large solids, abrasives,
and corrosives. They are self-priming and can run dry. Their versatility
makes them a good choice for pumping wastewater, acids, and foods. The
shortcomings of air-operated diaphragm pumps are that they require air
to operate (this may actually be a benefit if the pump is in an area where
compressed air is available but electricity is not), they have limitations
on flow and pressure, they produce fairly large pressure pulsations, and
they are quite energy inefficient. Some designs also have problems with
the air valves stalling or freezing up, and some air valves require periodic
lubrication.
Another type of diaphragm pump, shown in Figure 1.27, is known as a
wobble plate pump. This pump type has the reciprocating action of several
pistons or diaphragms caused by a rotating plate mounted eccentrically on
the shaft. Advantages of this pump type include quite high pressure capability, sealless pumping, self-priming, and the capability of running dry or
with a blocked inlet line. Disadvantages include relatively low flows and
many moving parts.
Introduction to Pumps
35
FIGURE 1.27
Wobble plate pump. (Courtesy of Wanner Engineering, Inc., Minneapolis, MN.)
14. Miniature PD Pumps
There are several special classes of miniature PD pumps used for very low
flow rates that are required for many specialty original equipment manufacturer (OEM) machines such as lasers and kidney dialysis machines. OEMs
require low-cost and small-sized pumps to remain competitive in their marketplace. Several types of small pumps have evolved to support these needs.
The two major types of miniature PD pumps are described below.
a. Gear Pumps
Miniature gear pumps (Figure 1.28) are quite common, especially when
moderate pressures (typically less than 150 psi) are required. One of the
most common in this category consists of stainless steel construction, composite nonmetallic gears/bushings, combined with magnetic drive (sealless construction, discussed more thoroughly in Chapter 5, Section V). These
pumps are typically rated from 5 mL/min up to 10 L/min. One of their
most important features is pulseless flow. They are ideal for smooth flow of
a wide variety of chemicals and coolants. Some of the more common applications include commercial ink jet printers, kidney dialysis machines, critical heating and cooling processes, laboratory instrumentation, dispensing
equipment, etc.
36
Pump Characteristics and Applications
Magnetic coupling
Motor
Bearings
Gears
FIGURE 1.28
Miniature gear pump. (Courtesy of Diener Precision Pumps, Lodi, CA.)
These pumps are typically driven by small AC or DC motors. Brushless
DC motors are rapidly becoming more popular, due to the safety associated
with low voltage and reliability. These pumps are suitable for high-speed
operation to keep the package size small, but life will be shorter.
b. Piston Pumps
An alternative for equipment manufacturers are miniature piston pumps
(Figure 1.29). They are preferred for high repeatability (<1%) and for applications where linearity with speed is important. Miniature piston pumps tend
to serve these needs nicely if the process can live with flow pulsations.
Cylinder
Piston
Motor
Seal
FIGURE 1.29
Miniature piston pump. (Courtesy of Diener Precision Pumps, Lodi, CA.)
Introduction to Pumps
37
Miniature piston pumps are generally constructed of nonmetallic materials such as ceramics, PEEK, Teflon, and PVDF. One of the most common
technologies, shown in Figure 1.29, adds a rotating motion to the reciprocating motion in such a way that no inlet and outlet check valves are needed.
Flow rate can be changed by changing the piston stroke and/or speed. They
are frequently driven by a stepper motor for very high positional accuracy,
which results in accurate flow or dispense capability.
This type of pump is used for pH balance in applications such as dialysis
equipment, laboratory instruments, and pesticide metering.
2
Hydraulics, Selection, and Curves
I. Overview
This chapter covers the basics of pump hydraulics, particularly as they apply
to centrifugal pumps. It shows the reader, step by step, how to determine
and analyze the criteria to completely select a centrifugal pump. Example
problems illustrate what information is needed to size a pump, how to go
about picking a pump from manufacturers’ catalogs, and how to determine
the required driver size.
NPSH (net positive suction head) and cavitation are described in great detail,
along with a number of examples. NPSH is one of the least-understood principles of pump hydraulics and is the cause for a great many pump problems. It is also often mistakenly blamed for other unrelated pump problems
that nevertheless have similar symptoms, such as air in a pump system or
misalignment.
Specific speed is an index normally used only by pump designers. However,
a better understanding of pump specific speed enables engineers and users
to see why certain pumps are used in certain applications, why centrifugal
pump curve shapes vary so dramatically, and why pump efficiencies vary so
widely from one type of pump to another.
An understanding of how the affinity laws work allows a user or engineer
to take a pump at one speed and figure out how performance changes when
it is run at another speed. Or, conversely, the affinity laws help predict the
required running speed to obtain a particular performance from a pump
whose performance at another running speed is known.
A proper understanding of system head curves is very important in determining where on the pump’s performance curve the pump operates. The
pump runs where the system tells it to operate, and the system head curve is
a good analytical tool to help determine how the system responds to changes
in flows, levels, and pressures. A better understanding of how to generate a
system head curve, and the knowledge of how the system head curve can
vary over the life of the pump, will help prevent many field problems, give a
longer life to the pump, and save energy.
1
2
Pump Characteristics and Applications
This chapter shows why and how multiple pumps perform together in a
system, arranged either in parallel or in series.
Finally, there is some discussion of the consequences of using excessive
“fudge factors” in sizing centrifugal pumps.
The initial decision that must be made in applying a pump is the decision regarding the type of pump to use. First, a decision must be made as
to whether the pump should be of the centrifugal or positive displacement
(PD) type. A PD pump is chosen if one or more of the key application criteria introduced in Chapter 1, Section VI.B, is present. Table 1.1 and Chapter
1, Section VI.C, can then be used to determine the PD pump alternatives to
consider.
If none of these application criteria are present, or if they are present but
are determined to not be a compelling reason to use a positive displacement
pump, then a centrifugal pump is chosen.
The next decision is to choose the type of configuration of centrifugal
pump to use (e.g., end suction, inline, split case, vertical turbine, submersible).
Unfortunately, making this decision is not a simple process and requires a
full understanding of the application criteria and system constraints, as well
as the centrifugal pump configuration options that are available. A great
deal of this decision-making process improves with experience, and one
only gets better at it by studying various types of pumping equipment and
learning more about each type’s benefits and shortcomings. Chapters 4, 5,
and 7 should prove helpful in this regard.
Sometimes, special design criteria or standards apply, such as American
National Standards Institute (ANSI), American Petroleum Institute (API),
nuclear standards (ASME Section III), or sanitary (FDA) requirements.
Sometimes, the pump configuration is dictated by the unique layout of the
pumping system (e.g., a borehole well, which requires a vertical turbine or vertical submersible pump) or special requirements of the liquid to be pumped
(e.g., a highly abrasive slurry). Any one of these requirements places limits on
the configurations of pumps that are available and on the number of pump
manufacturers that have a pump to offer.
For a great many applications, a “traditional” or “historical” configuration
is used (i.e., “the way we’ve always done it”). With the continuing evolution
of pump technology, the traditional approach is not necessarily always the
best option.
Sometimes, an economic analysis must be carried out to estimate total
lifetime costs to determine the type of pump configuration to use. The cost
items to be examined include first costs for the equipment, control system,
and the structure in which the equipment is housed; energy costs; and maintenance costs. Chapter 6 includes a more in-depth discussion of pump lifecycle costs.
As an example of the above, consider an application involving emptying a sump of liquid which is oily, corrosive, and abrasive. The choices of
pumps might include an end suction horizontal pump mounted above the
Hydraulics, Selection, and Curves
3
pit, a submersible pump mounted in the pit, or a vertical column sump
or vertical turbine pump partially submerged in the liquid. Each of these
is a unique pump configuration, with advantages and disadvantages for
the application, differing costs for the equipment, differing expected
maintenance concerns, differing operating costs, and differing costs for
construction of the structure. This is only one pump application problem, and virtually every application is unique with regard to the pumped
liquid’s properties, the hydraulic conditions, and the relative importance
the user places on first cost, energy costs, and maintenance expenses. The
material contained in Chapter 4, where each of the major pump types is
discussed as to its strengths and shortcomings, assists in this initial process of deciding the configuration of the pump type to use. Chapter 6 has
several examples of an economic analysis that can be made in comparing
several configuration alternatives.
Much of this chapter deals with the system in which a centrifugal pump
operates and the interaction of the pump and system. Fundamentally, there
are two types of systems, open systems and closed systems. An open system
(see Figure 2.9) is best exemplified by a transfer application, where a pump
is moving liquid from one vessel to another. In a closed system (see Figure
2.33), the liquid returns to the starting point, such as in a cooling or heating
circulation loop.
II. Pump Capacity
Chapter 1 introduced the characteristic head–capacity (H–Q) curve of a
centrifugal pump (see Figure 1.6). The two parameters that must be determined to size a pump are the capacity and the total head. Capacity is usually
expressed in gallons per minute (gpm), or for larger pumps, in cubic feet
per second (cfs) using USCS units, and in cubic meters per second, liters per
second, or cubic meters per hour using SI units. Total head, sometimes simply
referred to as head, is abbreviated TH or H, and is measured in feet (USCS
units) or meters (SI units).
The required capacity of the pump is normally dictated by the requirements of the system in which the pump is located. A process system is
designed for a particular throughput. A vessel must be filled (or emptied) in
a certain amount of time. An air conditioning system requires a particular
flow of chilled water to do the job it is designed to perform.
Regardless of the pump system being designed, it is usually possible to
arrive at a design flow rate for the pump. Sometimes there will be a planned
duty cycle for the pump that will require it to operate at only a fraction of the
full design capacity during certain periods and at higher or lower capacities at other times. Examples of this include a process plant with a variable
4
Pump Characteristics and Applications
capacity, or a chilled water system designed to meet a variable air conditioning load. The operating duty cycle of the pump should be estimated, using
the best available process and operations estimates, as this helps in selecting
the best type of pump and control system.
III. Total Head
As introduced in Chapter 1, Section V a centrifugal pump develops head by
raising the velocity of the liquid in the impeller, and then converting some
of this velocity into pressure in the volute or diffuser casing by a diffusion
process. The amount of head developed in the impeller is approximately:
H=
V2
2g
(2.1)
where V is the velocity at the tip of the impeller (ft/s) and g is the acceleration
of gravity (32.2 ft/s2).
The velocity at the impeller tip can also be expressed as
V=
rpm × D
229
(2.2)
where rpm is the pump speed (in revolutions per minute) and D is the impeller
diameter (in inches).
Substituting Equation 2.2 above for V in Equation 2.1 results in the following expression:
H=
(rpm × D)2
3.377 × 106
(2.3)
Equation 2.3 shows that the head developed by a centrifugal pump is only a
function of rpm and impeller diameter, and is not a function of specific gravity of the liquid being pumped. A pump moving a liquid up a static distance of
100 ft always has a required head of 100 ft (ignoring friction for the moment),
regardless of the specific gravity of the liquid. If the liquid were water (SG =
1.0), a pressure gauge located at the pump discharge would show a pressure,
using Equation 1.1, of 100 × 1.0/2.31 = 43.3 psig. If the liquid being pumped
up the 100-ft height were oil, with a specific gravity of 0.8, the gauge would
read 100 × 0.8/2.31 = 34.6 psig. The point of this is that the pump discharge
pressure expressed in psi (or equivalent metric units such as kPa) varies with
the specific gravity of the liquid, while the head expressed in feet (or meters)
of liquid remains constant for liquids of different density. This is why one
Hydraulics, Selection, and Curves
5
should always convert pressure terms into units of feet (or meters) of head
when dealing with centrifugal pumps. A pump’s head–capacity curve does
not require adjustment when the specific gravity of the liquid changes. On the
other hand, as will be demonstrated shortly, the horsepower curve does vary
with varying specific gravity. Further, as illustrated in Chapter 3, Section II,
the head–capacity curve is also affected by the viscosity of the pumped liquid.
To determine the required size of a centrifugal pump for a particular
application, all the components of the system head for the system in which
the pump is to operate must be added up to determine the pump total head
(TH). There are four separate components of total head, and they refer back
to the discussion in Chapter 1 on the reasons for raising the pressure of a
liquid. The four components of total head are
1. Static head
2. Friction head
3. Pressure head
4. Velocity head
Each of these four components of head must be considered for the system in
which the pump is to operate, and the sum of these is the total head (TH) of the
pump. Note that the last of these components, velocity head, may or may not
need to be included in the calculation of the components of system head for
sizing a pump, depending on the point of reference for the calculation (where
the pressures for pressure head and the levels for static head are measured).
The four components of system head are each discussed separately below.
A. Static Head
Static head is the total elevation change that the liquid must undergo. In most
cases, static head is normally measured from the surface of the liquid in the
supply vessel to the surface of the liquid in the vessel where the liquid is
being delivered. The total static head is measured from supply vessel surface
to delivery vessel surface, regardless of whether the pump is located above
the liquid level in the suction vessel (which is referred to as a “suction lift”)
or below the liquid level in the suction vessel (“suction head”). Figure 2.1
shows an example of a pump with a static suction head, and defines static
suction head, static discharge head, and total static head. Figure 2.2 shows a
system with a static suction lift.
Figure 2.3 shows that when the discharge piping goes above the delivery
vessel liquid surface, such as in the spray head shown, part of the static head
that would be measured to the high point in the piping is recovered due to
the siphon effect in the final downhill leg of pipe. Note that the static head
to the high point of the piping in Figure 2.3 must be considered when initially filling the system, to ensure that the pump develops enough head to
6
Pump Characteristics and Applications
Total
static
head
Static
discharge
head
Static
suction
head
FIGURE 2.1
Static suction head, static discharge head, and total static head.
reach the high point of the system. Figure 2.4 shows a submerged return line,
where the reference for static head reverts back to the delivery vessel liquid
surface.
Note that for a pump in a closed loop system (one where the return line
goes directly back to the pump suction, such as Figure 2.33), the total static
Static
discharge
head
Total
static
head
Static
suction
lift
FIGURE 2.2
Static suction lift, static discharge head, and total static head.
7
Hydraulics, Selection, and Curves
Total
static
head
Static
discharge
head
Static
suction
head
FIGURE 2.3
Recovered static head due to siphon effect moves reference point for static head to outlet of pipe.
Total
static
head
Static
suction
head
Static
discharge
head
FIGURE 2.4
Submerged return line moves reference point for static head to delivery vessel liquid surface.
head is zero, since all of the static head from the pump to the high point in
the piping system is recovered in the downhill leg of the system. As was the
case regarding Figure 2.3, the static head to the high point of the closed system must be considered when the system is initially filled, to ensure that the
pump develops enough head to overcome the static head to the high point of
the system during the initial fill.
In fully closed systems, since the liquid is not vented to atmosphere and
is undergoing temperature changes, the system should include a thermal
expansion tank at the suction of the pump. This tank, which usually includes
an air side separated from the water by a rubber diaphragm, protects the
8
Pump Characteristics and Applications
piping system from overpressurizing due to thermal expansion of the liquid,
and it also establishes a reference point for measuring net positive suction
head (described in Section VI to follow). Making an adjustment of the pressure on the air side of the thermal expansion tank (if the thermal expansion
tank is one where such an adjustment is permitted) allows one to change
the local pressure at every point in the system. In some systems the thermal
expansion tank may be a simple vented tank, located at the high point in the
system.
If the pressure head requirements of the system (see Section III.C below)
are given by gauge pressure readings at some point in the system suction
and discharge piping rather than at the supply and delivery vessels, then
the value of static head is the difference in elevation of the pressure gauges,
rather than the difference in elevation between the liquid surface in the supply and delivery vessels.
B. Friction Head
Friction head is the head necessary to overcome the friction losses in the piping, valves, fittings, and components such as heat exchangers for the system
in which the pump operates. Friction loss in a piping system varies as the
square of the liquid’s velocity (assuming fully turbulent flow). The smaller
the size of the pipe, valves, and fittings for a given flow rate, the greater the
friction head loss. In designing a piping system, if smaller sizes of pipes,
valves, and fittings are chosen, the installed capital cost of the piping system
is reduced. However, the tradeoff is that this results in higher total pump
head due to the increased friction head loss. This, in turn, increases pump
and driver capital cost and also increases lifetime energy costs. This relationship between capital costs and pumping costs is illustrated in Figure 2.5,
Costs
Total costs
Pumping costs
Piping capital costs
Optimum pipe
diameter
Pipe diameter
FIGURE 2.5
Piping capital costs and pumping costs both vary with pipe diameter, but in opposite ways.
Hydraulics, Selection, and Curves
9
which shows that an evaluation of several alternatives may produce an optimum pipe size selection. Another point is that choosing smaller suction lines
might cause the pump to cavitate due to the increased suction line friction
losses, as is discussed in more detail in Section VI to follow.
In theory, friction losses that occur as liquid flows through a piping system
must be calculated by means of complicated formulae, taking into account
such factors as liquid density and viscosity, and pipe inside diameter and
material. Luckily, these formulae have been reduced to tables and charts that,
although somewhat tedious and repetitive, are nevertheless not too complex.
Table 2.1 shows a typical pipe friction table for water at 60°F flowing through
schedule 40 steel pipe [2]. If the pumped liquid is other than water, or if the
pipe schedule or material is other than schedule 40 steel pipe, a different
table or an adjustment to the table must be used. Actually, depending on the
degree of accuracy required, Table 2.1 may often be used with other aqueous
(low viscosity) liquids and with different piping materials or schedules, at
least for preliminary sizing. Friction data for other pipe materials and inside
diameters are often found in engineering data tables or are sometimes available from pipe manufacturers. Some tables are available for liquids other
than water or computer software data tables can be used. (Refer to Chapter 3
Section III, for further discussion of the benefits of using computer software
for determining friction head.)
To determine the friction head for a particular section of piping, enter Table
2.1 at the planned design capacity and choose a line size. Usually, velocity of
the fluid, in feet per second (labeled “V” in Table 2.1), is used as the criterion
for choosing at least a preliminary line size, with the trade-off between piping system cost, pump capital cost, and lifetime energy costs (as illustrated
in Figure 2.5) being considered. Common velocity guidelines are 4 to 6 ft/s
for suction piping and 7 to 10 ft/s for discharge piping. The recommended
range can vary widely, depending on the particular application. Important
factors to consider when establishing the recommended velocity range for a
particular application include the abrasiveness of the liquid, pipe material,
and settling properties of any suspended solids.
Note that some engineers mistakenly believe that the suction and discharge
connection sizes of the pump they are using or considering to use for a particular application must be the same sizes as those chosen for, respectively,
the suction and discharge piping. This is quite often not the case. The pump
is designed by the manufacturer to be as compact as possible for a given
hydraulic duty point, which sometimes results in velocities in the suction
and discharge connections of the pump that are higher than the velocities
that are the optimal sizes to be used in the suction and discharge lines (taking into account considerations of avoiding cavitation by improper suction
pipe sizing and avoiding increased pumping costs due to undersizing the
discharge piping). It is quite often the case, therefore, that the suction pipe
in a pump system is one or more sizes larger than the suction connection on
the pump and that the discharge pipe is one or more sizes different from the
316
10.6
9.50
14.8
4.39
258
8.45
7.39
6.34
5.28
4.75
4.22
3.70
3.17
2.64
2.11
1.58
1.06
3.40
2.49
1.73
1.40
1.11
0.849
0.624
0.433
0.351
0.277
0.212
0.156
0.108
0.069
0.039
0.017
V2
(ft)
2g
183
136
95.9
78.3
62.7
48.7
36.5
25.8
21.3
17.1
13.3
10.0
7.16
4.78
2.85
1.86
hf
(ft/100 ft)
1/2 in (0.622 in I.D.)
V (ft/s)
12.7
16.8
10
3.56
205
158
118
83.5
68.4
54.9
42.6
31.8
22.6
15.0
8.93
4.30
2.89
1.74
hf
(ft/100 ft)
14
15.1
9
2.81
2.15
1.58
1.10
0.889
0.702
0.538
0.395
0.274
0.176
0.099
0.044
0.028
0.016
V2
(ft)
2g
3/8 in (0.493 in I.D.)
12
13.4
8
8.40
11.8
398
7.56
6.72
5.88
5.04
4.20
3.36
2.52
1.68
1.34
1.01
7
3.69
326
259
200
148
105
69.0
40.1
19.1
12.7
7.6
3.7
V (ft/s)
10.1
15.42
5
2.99
2.36
1.81
1.33
0.923
0.591
0.332
0.148
0.095
0.053
0.024
hf
(ft/100 ft)
6
13.87
6.17
4.5
324
4.62
12.33
1.98
188
3.08
2.47
4.0
11.3
2.0
1.12
87.0
57.4
10.79
8.48
1.5
0.495
0.317
1.85
3.5
5.65
1.0
33.8
1.23
9.25
4.52
0.8
0.178
2.72
16.2
3.0
3.39
0.6
0.079
0.020
7.71
2.26
2.5
1.13
0.4
V (ft/s)
V2
(ft)
2g
hf
(ft/100 ft)
V2
(ft)
2g
V (ft/s)
0.2
U.S. gal/​
min
1/4 in (0.364 in I.D.)
1/8 in (0.269 in I.D.)
Pipe Friction: Water/Schedule 40 Steel Pipe
TABLE 2.1
14
12
10
9
8
7
6
5
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.8
0.6
0.4
0.2
U.S. gal/
min
10
Pump Characteristics and Applications
15.1
18.1
25
30
5.06
3.54
2.25
187
134
86.1
70.3
11.1
9.29
7.42
6.68
5.94
21.5
25.7
100
120
140
19.3
15.0
12.9
10.7
9.66
8.58
7.52
6.44
5.37
4.29
3.86
3.43
3.00
2.57
2.15
1.93
1.72
1.50
1.29
90
283
209
146
119
95.0
73.3
54.6
37.4
25.1
20.6
16.5
12.8
9.62
6.86
5.65
4.54
3.56
2.68
1.93
1.29
10.3
7.15
5.79
4.58
3.50
2.57
1.79
1.45
1.14
0.879
0.644
0.448
0.286
0.232
0.183
0.140
0.103
0.071
0.058
0.046
0.035
0.026
V2
(ft)
2g
197
138
112
89.2
68.8
51.0
36.0
29.5
23.5
18.5
13.6
9.66
6.34
5.22
4.20
3.28
2.48
1.77
1.46
1.18
0.93
0.70
hf
(ft/100ft)
1 1/4 in (1.3880 in I.D.)
V (ft/s)
17.2
10.5
7.71
5.35
4.33
3.43
2.62
1.93
1.34
0.857
0.694
0.548
0.420
0.308
0.214
0.173
0.137
0.105
0.077
0.053
0.034
hf
(ft/100ft)
80
26.0
12.0
20
1.82
56.3
5.20
4.45
70
10.8
18
1.44
43.5
32.6
22.3
9.63
16
1.10
0.810
3.71
60
8.42
14
23.0
3.34
18.6
7.22
12
0.563
18.8
2.97
50
6.02
10
0.456
15.0
2.60
16.7
5.42
9
0.360
11.8
2.23
45
4.81
8
0.276
8.87
1.86
1.48
14.8
4.21
7
0.203
6.32
4.21
40
3.61
6
0.141
0.090
13.0
3.01
35
2.41
5
V (ft/s)
V2
(ft)
2g
hf
(ft/100ft)
V2
(ft)
2g
V (ft/s)
4
U.S. gal/
min
1 in (1.049 in I.D.)
3/4 in (0.824 in I.D.)
22.1
18.9
15.8
14.2
12.6
11.0
9.46
7.88
7.10
6.30
5.52
4.73
3.94
3.15
2.84
2.52
2.21
1.89
1.58
1.42
1.26
V (ft/s)
7.56
5.56
3.86
3.13
2.47
1.89
1.39
0.965
0.783
0.618
0.473
0.347
0.241
0.154
0.125
0.099
0.076
0.056
0.039
0.031
0.025
V2
(ft)
2g
119
88.3
62.2
51.0
40.5
31.3
23.2
16.4
13.5
10.8
8.38
6.26
4.50
2.94
2.42
1.96
1.53
1.16
0.83
0.69
0.56
hf
(ft/100ft)
1 1/2 in (1.610 in I.D.)
(continued)
140
120
100
90
80
70
60
50
45
40
35
30
25
20
18
16
14
12
10
9
8
7
6
5
4
U.S. gal/
min
Hydraulics, Selection, and Curves
11
28.7
300
12.8
11.1
146
128
111
20.1
18.4
17.4
16.1
34.7
800
1000
30.4
21.7
17.4
15.2
13.0
12.2
11.3
10.4
9.55
8.68
7.81
6.94
6.08
5.21
4.34
3.47
2.60
2.17
700
160
103
79.2
58.5
51.3
44.5
38.1
32.2
26.7
21.9
17.4
13.5
10.0
7.11
4.66
2.72
1.94
1.28
1.00
0.75
V (ft/s)
26.0
17.4
11.2
8.55
6.28
5.47
4.72
4.02
3.38
2.79
2.26
1.79
1.37
1.00
0.698
0.447
0.251
0.174
0.112
0.085
0.063
hf
(ft/100ft)
600
33.5
26.8
280
9.60
95.0
14.7
500
24.9
260
8.18
80.0
13.4
26.8
22.9
240
6.88
66.3
12.1
10.7
9.38
8.04
6.70
5.36
4.02
3.35
2.68
2.35
2.01
V
(ft)
2g
2
2 1/2 in (2.469 in I.D.)
V (ft/s)
400
21.0
220
5.68
54.1
43.0
33.2
24.7
17.4
11.4
6.59
4.67
3.10
2.42
1.82
hf
(ft/100ft)
23.5
19.1
200
4.60
3.64
2.78
2.05
1.42
0.909
0.511
0.355
0.227
0.174
0.128
V
(ft)
2g
350
17.2
180
9.56
100
15.3
7.65
80
160
5.74
60
11.5
4.78
50
13.4
3.82
40
140
3.35
120
2.87
35
V (ft/s)
30
U.S. gal/
min
2
2 in (2.067 in I.D.)
Pipe Friction: Water/Schedule 40 Steel Pipe
TABLE 2.1 (Continued)
18.7
14.3
10.5
7.32
4.68
3.57
2.63
2.29
1.98
1.69
1.42
1.17
0.948
0.749
0.574
0.421
0.293
0.187
0.105
0.073
V
(ft)
2g
2
131
101
74.8
52.5
33.9
26.3
19.2
16.9
14.7
12.6
10.7
8.90
7.28
5.81
4.51
3.37
2.39
1.57
0.92
0.66
hf
(ft/100ft)
3 in (3.068 in I.D.)
32.5
26.0
22.7
19.5
16.2
13.0
11.3
9.74
9.09
8.44
7.79
7.14
6.49
5.84
5.19
4.54
3.89
3.25
2.60
1.95
V (ft/s)
16.4
10.5
8.02
5.89
4.09
2.62
2.00
1.47
1.28
1.11
0.943
0.792
0.655
0.530
0.419
0.321
0.236
0.164
0.105
0.059
V2
(ft)
2g
96.4
62.3
48.0
35.6
25.0
16.2
12.4
9.26
8.11
7.04
6.04
5.12
4.27
3.50
2.80
2.18
1.64
1.17
0.77
0.45
hf
(ft/100ft)
3 1/2 in (3.548 in I.D.)
1000
800
700
600
500
400
350
300
280
260
240
220
200
180
160
140
120
100
80
60
50
40
35
30
U.S. gal/
min
12
Pump Characteristics and Applications
4.84
6.32
8.00
9.87
14.2
19.3
17.6
20.2
22.7
25.2
30.2
35.3
600
700
800
900
1000
1200
1400
19.2
16.0
14.4
12.8
11.2
9.62
8.02
5.76
4.00
3.24
2.56
1.96
1.44
0.999
0.811
8.28
10.2
25.7
4000
5.01
23.1
59.9
18.0
3.68
2.56
2.07
1.64
1.25
0.920
0.639
0.518
0.409
0.313
0.230
0.160
0.129
0.102
V2
(ft)
2g
3600
19.6
35.5
3200
46.1
15.4
12.8
11.5
10.3
8.98
7.70
6.41
5.77
5.13
4.49
3.85
3.21
2.89
2.57
V (ft/s)
6.55
15.0
31.1
2800
34.2
23.8
19.4
15.4
11.8
8.76
6.17
5.05
4.03
3.13
2.34
1.66
1.37
1.09
0.90
0.72
0.56
0.42
0.30
hf
(ft/100ft)
22.6
18.4
14.5
11.2
8.31
5.86
4.79
3.82
2.95
2.20
1.56
1.27
1.02
0.80
0.60
0.42
0.35
0.28
hf
(ft/100ft)
8 in (7.981 in I.D.)
20.5
11.0
26.6
6.21
4.91
3.76
2.76
1.92
1.55
1.23
0.939
0.690
0.479
0.388
0.307
0.240
0.196
0.150
0.110
0.077
V
(ft)
2g
2400
20.0
17.8
15.5
13.3
11.1
9.99
8.88
7.77
6.66
5.55
5.00
4.44
4.00
3.55
3.11
2.66
2.22
V (ft/s)
7.67
61.0
49.7
39.5
30.4
22.5
15.8
12.9
10.2
7.93
5.88
4.16
3.42
2.72
2.22
1.78
1.38
1.03
0.74
0.61
0.38
0.49
hf
(ft/100ft)
2
6 in (6.065 in I.D.)
22.2
97.6
72.0
50.2
40.8
32.4
25.0
18.6
13.0
7.23
16.0
3.55
15.1
500
10.5
0.639
32.1
2.47
12.6
450
6.41
0.518
2000
2.00
11.3
400
8.47
5.77
0.409
12.9
1.58
10.1
6.92
5.13
0.313
28.8
1.28
9.07
360
5.51
4.49
0.230
1800
1.01
8.06
320
4.30
3.85
0.160
0.129
10.2
0.774
7.06
280
3.21
3.21
2.89
25.7
0.569
6.05
240
2.27
1.86
0.078
0.102
V
(ft)
2g
1600
0.395
5.04
200
2.25
2.57
V (ft/s)
7.83
0.320
4.54
180
1.16
1.49
hf
(ft/100ft)
2
5 in (5.047 in I.D.)
22.5
0.193
0.253
V
(ft)
2g
3.53
4.03
V (ft/s)
160
140
U.S. gal/
min
2
4 in (4.026 in I.D.)
(continued)
4000
3600
3200
2800
2400
2000
1800
1600
1400
1200
1000
900
800
700
600
500
450
400
360
320
280
240
200
180
140
160
U.S. gal/
min
Hydraulics, Selection, and Curves
13
7.32
8.14
10.2
12.2
14.2
16.3
18.3
20.3
24.4
28.5
32.5
36.6
1800
2000
2500
3000
3500
4000
4500
5000
6000
7000
8000
9000
20.8
16.5
12.6
9.26
6.43
5.21
4.12
3.13
2.32
1.62
1.03
0.834
0.659
34.6
27.5
21.1
15.6
10.9
8.88
7.07
5.46
4.06
2.86
1.86
1.52
1.21
0.940
0.703
25.8
22.9
20.1
17.2
14.3
12.9
11.5
10.0
8.60
7.17
5.73
5.16
4.59
4.01
3.44
20,000
42.7
33.2
28.5
23.7
21.3
19.0
16.6
14.2
11.9
10.7
9.48
8.30
7.11
5.93
4.74
4.27
3.79
3.32
2.85
2.37
18,000
33.5
24.8
17.4
14.1
11.2
8.63
6.39
4.47
3.65
2.92
2.25
1.68
1.19
0.776
0.636
0.509
0.395
0.296
0.210
0.173
V (ft/s)
37.9
25.0
18.3
12.8
10.3
8.17
6.26
4.60
3.19
2.59
2.04
1.55
1.15
0.799
0.511
0.414
0.327
0.250
0.184
0.128
0.103
hf
(ft/100ft)
16,000
40.1
6.51
1600
0.504
0.370
2.87
34.4
5.70
1400
0.500
2.58
14,000
4.88
1200
0.257
0.410
0.328
V
(ft)
2g
2
12 in (11.938 in I.D.)
V (ft/s)
12,000
4.07
1000
0.208
0.165
hf
(ft/100ft)
28.7
3.66
V
(ft)
2g
10,000
3.25
900
V (ft/s)
800
U.S. gal/
min
2
10 in (10.020 in I.D.)
Pipe Friction: Water/Schedule 40 Steel Pipe
TABLE 2.1 (Continued)
28.3
22.4
17.1
12.6
8.74
7.08
5.59
4.28
3.14
2.18
1.77
1.40
1.07
0.786
0.546
0.349
0.283
0.224
0.171
0.126
0.087
V
(ft)
2g
2
33.9
26.8
20.7
15.2
10.7
8.7
6.90
5.32
3.95
2.78
2.27
1.81
1.40
1.04
0.738
0.483
0.395
0.317
0.247
0.185
0.131
hf
(ft/100ft)
14 in (13.124 in I.D.)
36.3
32.7
29.0
25.4
21.8
18.2
16.3
14.5
12.7
10.9
9.08
8.17
7.26
6.35
5.45
4.54
3.63
3.27
2.90
V (ft/s)
20.5
16.6
13.1
10.0
7.38
5.12
4.15
3.28
2.51
1.84
1.28
1.04
0.820
0.627
0.461
0.320
0.205
0.166
0.131
V2
(ft)
2g
21.2
17.2
13.5
10.4
7.69
5.38
4.38
3.49
2.69
2.01
1.41
1.15
0.921
0.718
0.535
0.377
0.248
0.203
0.163
hf
(ft/100ft)
16 in (15.000 in I.D.)
20,000
18,000
16,000
14,000
12,000
10,000
9000
8000
7000
6000
5000
4500
4000
3500
3000
2500
2000
1800
1600
1400
1200
1000
900
800
U.S. gal/
min
14
Pump Characteristics and Applications
11.5
14.3
17.2
20.1
22.9
25.8
28.7
31.6
34.4
37.3
40.2
43.0
8000
10,000
12,000
14,000
16,000
18,000
20,000
22,000
24,000
26,000
28,000
30,000
28.8
25.1
21.6
18.4
15.5
12.8
10.4
8.19
6.27
4.60
3.20
2.05
1.15
25.5
22.2
19.2
16.5
13.9
11.5
9.33
7.41
5.69
4.21
2.97
1.93
1.11
0.781
0.511
34.6
32.3
30.0
27.7
25.4
23.1
20.8
18.5
16.2
13.8
11.5
9.23
6.92
5.77
4.62
24.7
20.9
17.5
14.3
11.4
8.91
7.76
6.69
5.70
4.79
3.96
3.21
2.53
1.94
1.42
0.989
0.633
0.356
0.247
0.158
V2
(ft)
2g
24 in (22.624 in I.D.)
Hydraulic Institute, Engineering Data Book, 2nd edition. Parsippany, NJ, 1990; www.pumps.org. With permission.
39.9
50,000
Source:
36.7
30.3
27.1
23.9
22.3
20.7
19.2
17.6
16.0
14.4
12.8
11.2
9.58
7.98
6.38
4.79
3.99
3.19
46,000
23.2
18.7
14.6
12.7
11.0
9.39
7.91
6.56
5.35
4.26
3.29
2.44
1.70
1.11
0.645
0.455
0.298
0.174
V (ft/s)
33.5
29.9
23.9
18.6
16.2
14.0
11.9
10.0
8.28
6.71
5.30
4.06
2.98
2.07
1.32
0.745
0.517
0.331
0.186
hf
(ft/100ft)
42,000
43.9
8.61
6000
0.799
0.512
3.46
38,000
7.17
5000
0.139
0.297
39.2
5.74
4000
0.128
0.288
V (ft/s)
V2
(ft)
2g
hf
(ft/100ft)
V2
(ft)
2g
34,000
2.87
4.30
3000
V (ft/s)
2000
U.S. gal/
min
20 in (18.812 in I.D.)
18 in (16.876 in I.D.)
15.5
13.2
11.0
9.00
7.22
5.68
4.96
4.29
3.67
3.10
2.58
2.10
1.67
1.29
0.959
0.671
0.441
0.257
0.181
0.120
hf
(ft/100ft)
50,000
46,000
42,000
38,000
34,000
30,000
28,000
26,000
24,000
22,000
20,000
18,000
16,000
14,000
12,000
10,000
8000
6000
5000
4000
3000
2000
U.S. gal/
min
Hydraulics, Selection, and Curves
15
16
Pump Characteristics and Applications
discharge connection on the pump. This means, of course, that appropriately
sized piping reducers/expanders must be used at the pump inlet and outlet.
With the design capacity and the chosen preliminary pipe size, the friction tables give the head loss in feet per 100 linear feet of pipe (labeled “hf”
in Table 2.1). Note the two different uses of the term “foot” here, with Table
2.1 showing feet of head loss per 100 linear feet of pipe. This means that the
value found in Table 2.1 must be multiplied by the actual pipe length divided
by 100 to get the total friction head loss in a given length of pipe. This is
expressed by the formula:
H f (pipe) = h f ×
L
100
(2.4)
where L is the pipe length (in feet).
The friction loss in valves and fittings is given by the formula
H f (valve/fitting) = K ×
V2
2g
(2.5)
The value of V2/2g in Equation 2.5 for different flow rates and different valve/
fitting diameters is found in the pipe friction table (Table 2.1). The value of K,
the resistance coefficient for the particular valve or fitting, is determined using
one of the charts in Figure 2.6, which has a different K chart for each type of
valve or fitting [2]. To find K for a particular size of valve or fitting, take the
nominal size of the valve or fitting in inches, read up to the heavy line on the
chart, and then read on the left scale to get the K factor. For example, from
Figure 2.6, K for a 6-in flanged gate valve is 0.09. Multiplying the value of K
determined from Figure 2.6 times V2/2g gives the friction loss through the
particular valve or fitting, expressed in feet.
The resistance coefficient, K, for reducers or enlargers is a function of
how much the pipe diameter changes (smaller diameter divided by larger
diameter), and by how quickly the change occurs (indicated by the angle
of approach and the approach length). Refer to Ref. [1] for the formulae to
calculate the resistance coefficients for reducers and enlargers. Typically, a
reasonably sized reducer or enlarger has a K value of 0.3 to 1.0.
Note that the K values for valves shown in the Figure 2.6 charts are generic
only. If particular valves are already chosen, the valve manufacturer may
have more precise resistance coefficients.
Note that some valve manufacturers provide a flow coefficient, Cv, instead of
a resistance coefficient K. The mathematical relationship between these two
coefficients is given by the following formula:
K = 891d 4/C2v
where d is the inside diameter of the connecting piping (in inches).
(2.6)
17
Hydraulics, Selection, and Curves
Regular
screwed
45° ell.
Bell-mouth
inlet or reducer
K = 0.05
Long
radius
flanged
45° ell.
Inward projecting pipe
K = 1.0
1
Regular
flanged
90° ell.
K
Chart 1
2
1
4
Screwed
tee
.8
.6
Branch
flow
.4
.3
D .2
.3
2
1
.5
4
.6
Line
flow
2
4
6
10
20
K
.3
Flanged
tee
.2
D .1 1
Branch
flow
2
4
6
10
20
Where:
h = Frictional resistance in feet of liquid
V = Average velocity in ft/sec in a pipe of corresponding diameter
V²
h = K 2g
2
4
6
10
20
2
4
K 2
1
.5
1
Reg.
Long
radius
.1
1
2
4
6
10
20
2
4
1
K .8
.6
D .3
K
.5
.1
1
.3 .5
1
3
2
D
K .4
.3
.2
D .15 1
Long
radius
flanged
90° ell.
.5
Line
flow
4
2
.4
.3
.2
K
D
2
D .6
.3
Long
radius
screwed
90° ell.
.2
.1
D
1
.6
D .3
Flanged
return
bend
1
.5
K
.3
Screwed
return
bend
Note: K decreases with
increasing wall thickness of
pipe and rounding of edges
K
.4
.3
.2
D .3
Square edged inlet
K = 0.5
Regular
screwed
90° ell.
K .6
2
4
.2
K .1
.06
1
D
2
4
6
10
20
2
4
6
10
20
K .1
.6
D
.4
1
g = 32.17 ft/sec 2
K = Resistance coefficient for valve or fitting
FIGURE 2.6
Resistance coefficients (K) for valves and fittings. (From Hydraulic Institute, Engineering Data
Book, 2nd edition. Parsippany, NJ, 1990; www.pumps.org. With permission.)
18
Pump Characteristics and Applications
1.5
20
K
Basket
strainer
10
Screwed
D
6
.6
.3
1
2
K
D
4
1
.8
.6
.4
1
2
4
6
10
20
6
10
20
1
2
4
15
Globe
valve
K
Flanged
D
K
1.0
Foot .8
valve
.6
.4
D
10
6
4
1
2
4
6
10
20
1
2
4
.3
K
.2
Screwed
.1
D
.3
.6
2
1
4
.2
Gate
valve
K
Flanged
D
K
.1
.06
.04
.03
Couplings
and unions
4
2
1
6
10
D
20
Screwed 6
K 4
V
V
2
Swing
check
valve
1
D .3
K
Flanged
.6
1
2
4
6
10
20
1
D 1
K
2
4
6
V
4
.6
Reducing bushing
and coupling
V
h = K 2g
D
1
.3
.6
1
2
Sudden enlargement
2
4
6
10
if A → ∞ so that V = 0
h = V Feet of fluid
2g
4
4
3
2
1.5
1
D
V
h = (V — V) feet of fluid
2g
Flanged
Chart 2
.3
2
2
K
.04
.03
Used as reducer K = 0.05  2.0
used as increaser loss is up
to 40% more than that caused
by a sudden enlargement
3
Screwed
Angle
valve
.1
.08
.06
20
V
h = K 2g
FIGURE 2.6
(Continued) Resistance coefficients (K) for valves and fittings. (From Hydraulic Institute,
Engineering Data Book, 2nd edition. Parsippany, NJ, 1990; www.pumps.org. With permission.)
19
Hydraulics, Selection, and Curves
The K factors for valves in Figure 2.6 are based on fully open positioning
of the valves. For control valves or other valves that may be positioned other
than fully open, a value of K or Cv must be established for the valve in the
partially closed position. Valve manufacturers can help in this effort.
If the system pressure head requirements (see Section III.C below) are
given as gauge readings at some point in the system suction and discharge
piping, rather than at the supply and delivery vessels, then the friction head
consists only of friction losses in the portion of suction and discharge piping
located between the gauges.
C. Pressure Head
Pressure head is the head required to overcome a pressure or vacuum in
the system upstream or downstream of the pump. It is normally measured
at the liquid surface in the supply and delivery vessels. If the pressure in
the supply vessel from which the pump is pumping and the pressure in the
delivery vessel are identical (e.g., if both are atmospheric tanks), then there is
no required pressure head adjustment to TH. Likewise, there is no pressure
adjustment to TH for a closed loop system.
If the supply vessel is under a vacuum or under a pressure different than
that of the delivery vessel, a pressure head adjustment to TH is required.
The pressure or vacuum must be converted to feet. Pressure in pounds per
square inch converts to feet by Equation 1.1, rewritten as Equation 2.7:
feet = psi × 2.31/SG
(2.7)
where psi is pounds per square inch and SG is the specific gravity.
Vacuum, usually expressed in inches of mercury (in Hg) using USCS units,
is converted to feet of head by the formula:
Vac. (feet ) =
Vac. (in Hg) × 1.133
SG
(2.8)
Vacuum can also be expressed in inches of water, centimeters of mercury, or
other terms, and the conversion to feet or meters of head can be found in the
conversion data found in Appendix B at the end of this book.
If the suction vessel is under vacuum, the amount of vacuum (equivalent to
gauge pressure, converted to feet) must be added to the delivery vessel gauge
pressure (also converted to feet) to get the total pressure adjustment to TH.
(It is actually subtracting a negative value.) Alternatively, the vacuum in the
suction vessel can be expressed in absolute terms by subtracting the amount
of vacuum from the barometric pressure (both terms expressed in inches of
mercury). This value is then converted to feet using Equation 2.8, and the
result is subtracted from the absolute pressure in the delivery vessel (in feet).
20
Pump Characteristics and Applications
If the suction vessel is under positive pressure (but different from the pressure of the delivery vessel), then the suction vessel pressure (converted to feet)
should be subtracted from the delivery vessel pressure (converted to feet) to
get the pressure adjustment to TH. Either gauge pressures or absolute pressures can be used in this case, as long as the pressures in the supply and
delivery vessels are expressed consistently as one or the other. If the above
seems confusing, the example in Section IV to follow should help clarify this.
If the pressure requirements are given as gauge readings at some point in
the system suction and discharge piping, rather than at the liquid surface
in the supply and delivery vessels, then the value of static head (see Section
III.A above) is the elevation difference between the two gauges. The velocity
head (see Section III.D below) is taken at the points of the gauge connections, and the friction head (see Section III.B above) is only the friction losses
between the two gauges.
D. Velocity Head
Velocity head is the energy of a liquid as a result of its motion at some velocity V. The formula for velocity head is
Hv =
V2
2g
(2.9)
This value is found in Table 2.1, expressed in feet of head. The value of velocity
head is different at the suction and discharge of the pump, because the size of
the suction piping is usually larger than the size of the discharge piping.
Note that the normal procedure in sizing centrifugal pumps measures
the required pressure head at the liquid surface in the supply and delivery
vessel, as well as establishing static head values from these levels. In this
situation, because velocity is zero at the liquid surface in the supply and
delivery vessels, velocity head at these points is also zero. Velocity head is
only included in the calculation of required pump total head when the pressure head requirements are given as gauge readings at some point in the
system suction and discharge piping.
To determine the velocity head component of TH in those situations where
it is appropriate, it is necessary to calculate the change in velocity head from
suction to discharge. Examining Table 2.1, it is seen that in the typical pump
configuration, where the suction connection is one or two sizes larger than
the discharge connection, the change in velocity head across the pump is normally quite small, usually on the order of no more than 1 or 2 ft. With the
TH of many pumps being several hundred feet or more, many pump selectors choose to totally ignore the effect of velocity head because the change of
velocity head is often less than 1% of TH. However, considering the change of
velocity head to be negligible is not always a valid assumption, and a quick
Hydraulics, Selection, and Curves
21
check of the velocity head change using Table 2.1 is advisable before dismissing this term. The one situation where velocity head cannot be ignored is
when sizing a very low head pump. For example, if the TH of the pump being
sized is only, say, 15 ft, then a velocity head change of 1.5 ft represents 10% of
the total pump head. This is a significant amount and might affect the pump
size or impeller diameter that is required. As another example, consider a
pump system with a suction size of 3 in and a discharge size of 1.5 in, and
having a capacity of 120 gpm and a TH of 50 ft. Table 2.1 shows the change of
velocity head across this pump to be about 5.1 ft, or slightly over 10% of TH.
The 10-step procedure below summarizes the procedure for determining the
total head for a pump, using the methods and formulas covered in this section:
1. Choose an appropriate design velocity range for the suction and
discharge piping, based on the parameters discussed in Chapter 2,
Section III.B. If using relatively clear liquid and standard pipe materials, use suction piping design velocity range of 4 to 6 ft/s, and discharge piping design velocity range of 7 to 10 ft/s.
2. Choose the suction pipe size, based on the design velocity range. Use
the pipe friction tables (Table 2.1) to do that, finding the pipe size that
meets the design velocity range. If the piping already exists (i.e., if a
pump is being replaced in an existing system), ignore Step 1 and make
note of the existing suction pipe diameter. For the pipe size chosen,
make note of the values of V2/2g and hf found in the pipe friction tables.
3. Repeat Step 2 for the discharge pipe.
4. Determine the static head, Hs, which is usually equal to the change
in elevation from the normal liquid level in the suction supply vessel to the normal liquid level in the delivery vessel. The exception is
when the liquid free falls, as per Figure 2.3.
5. Calculate the friction head losses in the suction pipe. these include
• Pipe losses: This is the value of hf times the actual pipe length,
divided by 100.
• Fitting losses: For each fitting type, multiply the K value from the
appropriate chart in Figure 2.6 times V2/2g, times the quantity of
that type of fitting. Include the sudden contraction loss where the
suction pipe leaves the suction vessel (use a K of 0.5 times V2/2g
for square edge. If not square edge, see Figure 2.6).
• Valve losses: For each valve type, multiply the K value from the
appropriate chart in Figure 2.6 times V2/2g, times the quantity
of that type of valve.
6. Repeat Step 5 for the friction head losses in the discharge pipe.
Include the sudden enlargement loss where the discharge pipe enters
the delivery tank (equal to a K of 1 times V2/2g). Friction head, Hf, is
the sum of the friction head losses in the suction and discharge pipes.
22
Pump Characteristics and Applications
7. Calculate the pressure head, Hp, which is equal to 0 if the system is a
completely closed system, or if the system is open but the pressure in
the delivery vessel is the same as the pressure in the supply vessel. If
pressures in the supply and delivery vessels are different from each
other, they must be converted to feet of head. Use Equation 2.7 to convert pressure in psi to feet of head or Equation 2.8 to convert vacuum
in inches of mercury to feet of head. When using gauge pressures for
these calculations, the Pressure Head is equal to the converted feet
of head in the delivery vessel minus the converted feet of head in the
supply vessel if the supply vessel pressure is atmospheric pressure
or higher, and is equal to the converted feet of head in the delivery
vessel plus the converted feet of head in the supply vessel if the supply vessel pressure is below atmospheric pressure. The best way to
avoid confusion is to use absolute pressure in the supply and delivery vessels to convert to feet, rather than gauge pressures. If absolute
pressures are used for these conversions, pressure head is always
equal to the converted feet of head in the delivery vessel minus the
converted feet of head in the supply vessel.
8. Determine the velocity head, Hv. This value is 0 if the normal reference point for head calculations is used, namely the liquid surface levels in the suction and delivery vessels. If any other reference
points are used, calculate V2/2g at the reference point in the discharge pipe and subtract the value of V2/2g at the reference point
in the suction pipe. These values of V2/2g can be found in the pipe
friction table, Table 2.1, for the respective pipe sizes.
9. Add the values of all head losses together, including static head
(Step 4), friction head in the suction pipe (Step 5), friction head in the
discharge pipe (Step 6), Pressure Head (Step 7), and, if applicable,
Velocity Head (Step 8). The sum of all of these is the Total Head (TH).
10. Choose a pump to produce a head equal to, or slightly greater than
the calculated Total Head, at the desired flow rate.
IV. Performance Curve
Once the pump configuration and rating (capacity and head) have been
determined, as described in the three preceding sections, the next step in
the selection process is to decide which pump speeds should be considered.
It is quite often the case that two or more operating speeds may be commercially available for a particular pump rating and configuration. Each of
these speeds results in a different sized pump, each having different first
cost, operating cost, and maintenance cost. Section XIII covers the criteria
23
Hydraulics, Selection, and Curves
that should be considered to determine which operating speed to consider
for a particular application. An example containing some of these criteria is
given in Chapter 6.
The available motor speeds for standard alternating current (AC) electric
motors are based on the following formula, for 60-cycle current:
rpm = 7200/N (at frequency = 60 Hz)
(2.10)
where N is the number of poles.
Electric motors have an even number of poles, starting with two. The commercially available, constant-speed AC electric motors with 60 Hz electrical
supply are as shown in Table 2.2.
Designations for slower speed motors than shown in Table 2.2 would follow Equation 2.10. The actual operating speeds of motors are slightly less
than the values shown in Table 2.2, due to electrical slippage between motor
rotor and stator. Thus, the operating speed of a two-pole motor is 3450 to
3550 rpm, the speed for a four-pole motor is 1750 to 1780 rpm, etc.
For 50-cycle current, which is common in Europe and some other parts of
the world, Equation 2.10 is revised to
rpm = 6000/N (at frequency = 50 Hz)
(2.11)
where N is the number of poles.
Accordingly, with 50-cycle current supply, commercially available AC electric motor speeds are 3000 rpm, 1500 rpm, etc.
The manufacturer determines which speeds will be offered for each pump
type and size, based on a number of design and application considerations
(Section XIII). In general, the larger the pump impeller (and pump capacity),
the slower the pump runs. Also, certain types of applications such as abrasive slurries or paper stock require slower pump speed than clean services.
Once the pump speeds to be considered have been determined, a centrifugal pump selection can be made. However, if a variable-speed pumping
system is being considered for a system requiring a range of flow, additional
information in the form of a system head curve must be developed before a
TABLE 2.2
AC Electric Motor Speeds—60 Hz
N (No. Poles)
rpm
2
4
6
8
10
12
3600
1800
1200
900
720
600
24
Pump Characteristics and Applications
determination can be made as to the required pump speeds for the various
flow requirements. (See further discussion on this subject in Chapter 6.) For
the moment, it is assumed that a constant-speed pump is being selected.
With a constant pump speed, the head–capacity relationship for centrifugal pumps is depicted in Figure 1.6. Most centrifugal pumps, however, have
the capability to operate over an extended range of head and flow, by trimming, or cutting the impeller diameter from its maximum size down to some
predetermined minimum size. Thus, for a given pump speed, a centrifugal
pump produces an envelope of head–capacity performance, as illustrated
in Figure 2.7. The upper and lower boundaries of the envelope in Figure 2.7
are dictated by the maximum and minimum impeller diameters that the
manufacturer offers for a particular pump size. The right and left boundaries of the envelope are the maximum and minimum flows for each impeller
diameter, established by the manufacturer for the particular pump. This is
discussed more thoroughly in later sections.
The upper boundary of performance shown in Figure 2.7 is based on the
maximum impeller diameter that will physically fit inside the pump casing. The minimum impeller diameter offered by the manufacturer is based
on several criteria. Often, simple economics dictate the minimum impeller
diameter that the manufacturer offers. A point is reached at which further
trimming of impeller diameter makes for an uncompetitive offering, because
a competing pump manufacturer is likely able to offer a more competitive
pump with a smaller casing. Additionally, the efficiency of the pump usually
is lower if the impeller used is appreciably less than the maximum diameter.
Most pump manufacturers publish performance envelopes for an entire line
of pumps at a given speed, as illustrated in Figure 2.7. Once the pump capacity, head, configuration, and speed have been chosen, the preliminary selection of a pump size can be made from a family of envelope curves such as
shown in Figure 2.8. Once the pump sizes to be considered are chosen from
Maximum
impeller
diameter
H
(ft)
Minimum
impeller
diameter
Q (gpm)
FIGURE 2.7
Head–capacity envelope for a constant-speed centrifugal pump.
25
Hydraulics, Selection, and Curves
800
700
180
600
Total head — 3500 rpm (60 Hz)
220
140
100
60
20
0
100
30
20
10
0
150
500
600
1 1/2 × 3 – 13
1 × 2 –10
300
200
100
0
0 gpm
0 m3/h
310
270
230
200
20
40
200
3 × 4 – 13
(LTX)
100
500
70 80
700
100
1100
900
150
40
1400
0 0
200
250
= Scale change
Capacity — 3500 rpm (60 Hz)
60
200
100
140 160 200
300
500 600
800 1000 1200
300 400 500 600 700 800 1000 1400 1800 2200 2600 3400 4200 5000
200
4 × 6 – 17
6 × 8 – 17
6 × 8 – 15
8
180
×
10
160
140
–
15
120
140
60
80 100
140
180 220
300
500
Capacity — 1750 rpm (60 Hz)
240
200
G
3 × 4 – 13
2×
3–
8×
1 1/2 × 3 – 13
1
13
0–
4 × 6 – 13
120
13
1 1/2 × 3 – 10
100 1 × 2 –10
6 × 8 – 13
2× 3×4
–
3–
10
80
3 × 4 – 10H
2×
10
3–
4×6–10H
60 1 × 1
1 1/2 × 3 – 8
1/2–8 8
3
×
4
–8
7
40
–
2×3–6 × 4
20
3
4 × 6 – 10G
0
1 × 1 1/2 – 6 1 1/2 × 3 – 6
0 gpm 100
700
900 1000 1400 1800 2200 3000 3400
300 400 500
160
40
ft
1750/1450 rpm
STi
MTi/LTi
XLTi
0 3 20
m /h
80
4 ×6 –10G
300
60
300
3 × 4 – 8G
3× 4–7
160
Capacity —1450 rpm (50 Hz)
40
gpm 100
ft 0
350
/2
11
1 1/2 ×3 –6 2 × 3 – 6
100
500
m
400 120
3 × 4 – 10
2 ×3 –10
2 ×3 – 8
–8
×3
1 × 1 1/2 – 6
1100 ft
2 × 3 –13
(LTX)
1 1/2 × 3 – 10
1 ×1 1/2 – 8
250
800
STi
MTi/LTi
500
400
200
700
7
40
100
400
–1
50
80
300
10
75
60
65
60
8×
Total head — 1750 rpm (60 Hz)
95
50
200
3500/2850 rpm
0 m3/h 20
m
40
Total head — 2850 rpm (50 Hz)
ft
20
700
8
×
–
10
H
16
100
60
20
1000
= Scale change
55
45
40
30
80
40
5000 5800 6600 7400
m
75
0
20
10
Total head — 1450 rpm (50 Hz)
m
Capacity — 2850 rpm (50 Hz)
0 m3/h
0 gpm
0
1400
FIGURE 2.8
Typical family of envelope performance curves for a line of end suction centrifugal pumps,
shown at two-pole and four-pole speeds. (Courtesy of Goulds Pumps, Inc., a subsidiary of ITT
Corporation.)
26
Pump Characteristics and Applications
these envelope curves, the specific pump curves for those pump sizes can be
examined in more detail to ascertain information such as the required impeller diameter, efficiency, horsepower, etc. Note that the different shading of the
envelope curves shown in Figure 2.8 refer to different bearing frame sizes.
The convention in the United States for designating the size of a centrifugal pump is as follows:
Discharge size × suction size – maximum impeller diameter
All terms are expressed in inches. So, a centrifugal pump with a 6-in discharge flange, an 8-in suction flange, and a maximum impeller diameter of
15 in would be given a size designation of 6 × 8 – 15 (which is read “six by
eight – fifteen”). Note that the maximum impeller number is usually nominal.
The actual maximum impeller for a pump designated 6 × 8 – 15 may be 15 1/8
in. Some manufacturers occasionally have more than one impeller for a given
size casing, so the third digit may include a letter or other identification. Also
note that some manufacturers reverse the size notation, listing the suction size
first. So, the example pump mentioned above would be given the size 8 × 6 – 15
by some companies. Both notations are used in the industry, and because the
larger of the first two numbers is always the suction size (except wastewater
pumps, which often have the same size suction and discharge connections),
this should not cause confusion. Also note that some U.S. manufacturers use a
“×” symbol instead of a dash before the final number in the size designation.
The practice in Europe is to express the dimensions for the size designation
in millimeters. The size designation is most commonly either expressed in
the format first shown above as the standard in the United States (discharge
size × suction size – maximum impeller), or is simply shown as discharge—
maximum impeller diameter.
The following example serves to illustrate the important concepts that
have been introduced thus far in this chapter.
Example 2.1: TH Calculation/Pump Selection
PROBLEM
The piping for the system shown in Figure 2.9 must be sized, the pump
TH must be computed, and a centrifugal pump selected.
GIVEN
Capacity = 700 gpm
Liquid = Water at 60°F (SG = 1.0)
Pipes = Schedule 40 steel pipe, all values and fittings are flanged
Atm. press. = 14.7 psia = 29.9 in Hg
Pump configuration = end suction
Speeds to consider = 1800/3600 rpm (refer to Section XIII of this
chapter and also Chapter 6, Section II, for a discussion of speeds
to consider for a given application)
27
Hydraulics, Selection, and Curves
50 psig
50 ft
900 ft SCH 40
Steel pipe
5 in Hg
150 ft SCH 40
Steel pipe
FIGURE 2.9
System for example problem illustrating selection of line sizes, pump TH, and pump size
(Example 2.1); and NPSHa calculation (Example 2.5).
SOLUTION
Using Table 2.1, preliminary sizes of the suction and discharge piping
are chosen, based on a design velocity of 4 to 6 ft/s for the suction piping
and 7 to 10 ft/s for the discharge piping. Suction piping size is chosen as
8 in (velocity = 4.49 ft/s per Table 2.1) and discharge piping size is chosen
as 6 in (velocity = 7.77 ft/s per Table 2.1).
Note that pressure and static head requirements are given at the liquid
surfaces in the suction and delivery vessels. These are the usual reference points for pump TH calculation.
STATIC HEAD
The pump in Figure 2.9 is operating with a suction head. Static head, the
total change in elevation from suction vessel surface to delivery vessel
surface, equals 50.0 ft.
FRICTION HEAD
Using the selected line sizes, the line lengths shown on Figure 2.9, and
data from Table 2.1 and Figure 2.6, the pipe friction head is computed as
follows:
Suction:
Line loss
h f = 0.80 for 8 in pipe (Table 2.1)
H f = h f × L/100, from Equation 2.4
H f = 0.80 × 150/100 = 1.20 ft
28
Pump Characteristics and Applications
Valve loss
Quantity: 2 − 8 in flanged gate valves
K = 0.07 (Figure 2.6)
V 2/2 g = 0.31 (Table 2.1)
H f = K × V 2/2 g (Equation 2.5) × quantity
H f = 0.07 × 0.31 × 2 = 0.04 ft
Inlet (tank outlet) loss
K = 0.5 (Figure 2.6, using square edged inlet)
V 2/22 g = 0.31 (Table 2.1)
H f = K × V 2/2 g (Equation 2.5)
H f = 0.5 × 0.31 = 0.16 ft
Suction friction head = 1.20 + 0.04 + 0.16 = 1.4 ft
Discharge:
Line loss
h f = 3.13 for 6 in pipe (Table 2.1)
H f = h f × L/100, from Equation 2.4
H f = 3.13 × 900/100 = 28.17 ft
Valve loss
Quantity: 2 − 6 in flanged gate valves
K = 0.09 (Figure 2.6)
V 2/2 g = 0.94 (Table 2.1)
H f = K × V 2/2 g (Equation 2.5) × quantity
H f = 0.09 × 0.94 × 2 = 0.17 ft
Valve loss
Quantity: 1 − 6 in flanged check valve
K = 2.00 (Figure 2.6)
V 2/2 g = 0.94 (Table 2.1)
H f = K × V 2/2 g (Equation 2.5) × quantity
H f = 2.00 × 0.94 × 1 = 1.88 ft
Hydraulics, Selection, and Curves
Fitting loss
Quantity: 3 − Regular Flanged 90° Elbows
K = 0.28 (Figure 2.6)
V 2/2 g = 0.94 (Table 2.1)
H f = K × V 2/2 g (Equation 2.5) × quantity
H f = 0.28 × 0.94 × 3 = 0.79 ft
Sudden enlargement loss (at delivery tank)
h = V12/2 g (from Figure 2.6 for sudden enlargement)
h = 0.94 ft (Table 2.1)
Discharge friction head = 28.17 + 0.17 + 1.88 + 0.79 + 0.994
Discharge friction head = 31.95 ft
Total friction head = 1.40 ft + 31.95 ft = 33.4 ft
PRESSURE HEAD
Supply vessel pressure = 5 in Hg
Vac. (ft) = in Hg × 1..133/SG, from Equation 2.8
Vac. (ft) = 5 × 1.133 / 1.0 = 5.7 ft
Delivery vessel pressure = 50 psig
Head (ft) = psi × 2..31/SG, from Equation 2.7
Head (ft) = 50 × 2.31/1.0 = 115.5 ft
Total pressure head = 115.5 + 5.7 = 121.2 ft
Alternately, expressing both pressures in absolute terms,
using the given barometric pressure of 29.9 in Hg, the same
result is achieved.
Supply vessel = [(29.9 − 5) × 1.133]/ 1 = 28.2 ft
Delivery vessel = 64.7 psia = (64.7 × 2.31)/ 1 = 149.4 ft
Total pressure head = 149.4 − 28.2 = 121.2 ft
29
30
Pump Characteristics and Applications
VELOCITY HEAD
Because the pressure and static head requirements are referenced to the liquid surface in the suction and delivery vessels,
the velocity head at the liquid surfaces is zero.
TOTAL HEAD
TH = static + friction + pressure + velocity
TH = 50.0 + 33.44 + 121.2 + 0.0 = 204.6 ft
Using Figure 2.8, the preliminary pump selections are
1750 rpm: 6 × 8 − 15
4 × 6 − 17
3500 rpm: 3 × 4 − 8G
The individual curves for the preliminary selections should then be examined to evaluate each of the possible choices with regard to horsepower, efficiency, and NPSHr. See discussions on these subjects in Sections V, VI and
XIII to follow. Further analysis should then be done of first cost, power costs
to operate, and expected maintenance costs of the alternatives before a final
selection is made. Refer to Chapter 6 for an example of this further analysis.
Also, refer to the discussion in Chapter 3, Section III to learn how computer
software can be used to solve the above example problem.
V. Horsepower and Efficiency
Horsepower refers to the amount of energy that must be supplied to operate
a pump. An understanding of how to calculate horsepower and how to read
and interpret the horsepower data shown on the pump performance curve is
necessary to choose the correct size of driver for the pump. There are several
commonly designated expressions for horsepower. Water horsepower (WHP)
refers to the output of the pump handling a liquid of a given specific gravity,
with a given flow and head. The formula for WHP, in USCS units, is
WHP =
Q × H × SG
3960
(2.12)
where Q is the flow rate in gpm, H is the TH in feet, and SG is the specific gravity.
The constant 3960 is used when the units are as described in Equation 2.12.
The constant is obtained by dividing 33,000 (the number of ft-lb/min in one
31
Hydraulics, Selection, and Curves
horsepower) by 8.34 (the number of pounds per gallon of water). If Q is given
in cubic feet per second, Equation 2.12 becomes:
WHP =
Q × H × SG
8.82
(2.13)
Using SI units, the power WHP in watts is given by
WHP = 9797 × Q × H × SG
(2.14)
where Q is the flow rate (in m3/s) and H is the TH (in m).
If Q is given in liters per second, Equation 2.14 becomes
WHP = 9.797 × Q × H × SG
(2.15)
Brake horsepower (BHP) is the actual amount of power that must be supplied to the pump to obtain a particular flow and head. It is the input power
to the pump, or the required output power from the driver. The formula for
BHP, using the same units as Equation 2.12, is
BHP =
Q × H × SG
3960 × η
(2.16)
where η is the pump efficiency.
Other equations for BHP can be written using other USCS or SI units, by
taking Equations 2.13, 2.14, or 2.15, and adding pump efficiency, η, in the
denominator, as is done in Equation 2.16 above.
BHP is indicated on the pump performance curve as a function of pump
capacity, and is used to select an appropriate size of motor (or other driver
type) for the pump. Note that the BHP is a function of specific gravity. If the
pumped liquid’s specific gravity is other than 1.0, the BHP curve should be
adjusted accordingly, either by the manufacturer or by the engineer making
the motor selection.
Still another horsepower term that is used in studies and discussions
of pumping systems is wire-to-water horsepower. This term describes the
required power input into the driver, and is found by dividing BHP by the
motor efficiency. In the case of a pump using a variable-speed device or other
auxiliary driving equipment such as a gear box, BHP is divided by the combined efficiency of all of the driver components to obtain the wire-to-water
horsepower.
BHP is greater than WHP because of the fact that a pump is not a perfectly
efficient machine. There are actually four factors that cause a centrifugal
pump to be less than perfectly efficient, as described below.
32
Pump Characteristics and Applications
A. Hydraulic Losses
This term is a summary of internal losses in the impeller and volute or diffuser
due to friction in the walls of the liquid passageways and the continual change
of direction and recirculation of the liquid as it moves through the pump.
B. Volumetric Losses
This term refers to the leakage of a usually small amount of liquid from the
discharge side of a centrifugal pump to the suction side (the equivalent of
slip in a positive displacement pump). The liquid leaks past the wear rings
in a closed impeller pump and past the front edges of the vanes in an open
impeller pump. (Refer to Chapter 4, Section II.A) Other volumetric losses
occur between stages of multistage pumps, past some thrust balancing
devices (which are discussed in Chapter 4, Section II), and through leakage at seals and packing. Volumetric losses increase as internal clearances
are opened up due to wear and erosion in the pump. This causes the pump
to run less efficiently and increases BHP, as well as reducing flow and total
head that the pump produces.
C. Mechanical Losses
This term refers to the frictional losses that occur in the moving parts of
pumps that are in contact (bearings and packing or seals).
D. Disk Friction Losses
If the pump impeller is thought of as a rotating disk, rotating in very close
proximity to a fixed disk (the casing), there is a frictional resistance to this
rotation known as disk friction.
The pump efficiency is expressed as a decimal number less than 1, for example, 0.75 for 75% efficiency. The relative importance of the above four losses
varies from one pump type to another. Actual efficiencies for various types
of centrifugal pumps can vary widely, over a range from less than 30% to
over 90%, for reasons that are explained in more detail in Section XIII.
Comparing Equations 2.12 for WHP and Equation 2.16 for BHP, the only
difference between the two is the pump efficiency term. Therefore, the pump
efficiency is equal to the ratio of the two:
η=
WHP
Q × H × SG
=
BHP
3960 × BHP
(2.17)
The pump manufacturer uses Equation 2.17 to determine the pump efficiency at the time the factory pump performance test is done, as described
33
Hydraulics, Selection, and Curves
below. This same testing procedure can be done in the field as well to verify
pump performance and compare efficiency with the as new condition.
When a new pump is being designed by a pump manufacturer, there is
usually a predetermined objective for the pump’s flow and head at the best
efficiency point, as well as an expected maximum efficiency. However, it is
not until the performance test is run on the prototype that the performance
that the manufacturer lists in the catalog for that pump is finally determined.
In the early phases of the pump hydraulic design, the designer does calculations to determine the parameters of the design of the impeller and volute
or diffuser. These include selection of vane inlet angle, radius of curvature,
number of vanes, exit angle, etc.
Note that impeller and volute hydraulic design is beyond the scope of this
book. Readers interested in learning more about pump hydraulic design are
referred to Refs. [3] through [6] at the end of this book.
With the design parameters selected, the designer can then complete the
layout of the impeller and volute or diffuser. This allows creation of pattern
and machine drawings for these components. With the completion of the
design of the other mechanical components of the pump such as the stuffing
box and bearing assembly, a prototype pump can be built and made ready
for the performance test.
The manufacturer’s pump performance test is usually conducted with the
pump mounted on the floor above a large pit or sump filled with water. The
pump takes suction from the sump, the flow passes through instrumentation
that can measure flow Q and total head TH, and then the flow is returned
back to the sump. (Refer to Chapter 3, Section VI, for a more detailed discussion on how total head and flow are measured in the pump test.) The test
loop has a throttling valve to allow for variation of the flow and total head so
that the pump can be run over its full performance range.
Finally, the manufacturer’s laboratory test facility has the capability to
measure BHP, the power required by the pump. This is done in the laboratory in one of several ways. One common method measures the torque on
the shaft between the pump and motor, and converts this to horsepower by
the formula:
BHP =
rpm × T
5250
(2.18)
where T is the torque (in ft-lb).
Dynamometers are also used to measure torque. A more common approach
uses electrical instrumentation to measure the input power drawn by the
motor at a given flow rate, the wire-to-water horsepower previously discussed. This is then multiplied by the motor efficiency. The motor efficiency
is a value that is available from the motor manufacturer. (For most AC electric motors, the efficiency remains unchanged from full load to nearly 50% of
full load.) Motor input power times motor efficiency equals motor output
34
Pump Characteristics and Applications
power, and the motor output power thus measured is the pump BHP. This
is also the approach that would be used to measure BHP in a field test of a
pump, as described in Chapter 3, Section VI.D.
The pump is turned on in the test loop and the throttle valve is set at an
arbitrary position. Then, using the laboratory instrumentation, the values of
Q and TH are measured, as is BHP using one of the above-described methods. Then, using Equation 2.17 (with SG = 1.0 because the test loop contains
water), the value of pump efficiency η is determined for that particular point
on the pump curve. The data obtained from this test point (Q, TH, BHP, η)
are recorded, and then the throttle valve is repositioned and a new set of data
points is taken. This procedure is repeated over the full range of performance
of the pump. Usually, a minimum of five to seven points on the pump curve
are measured. The data can then be plotted to create the head–­capacity, BHP,
and efficiency curves for the pump, using a full-diameter impeller.
Figure 2.10 illustrates typical performance curves generated by the performance test just described. In a typical centrifugal pump, the head–capacity
90
Head (ft)
120
Shutoff
15.75 in
NPSHr (ft)
50
40
Preferred operating
region
60
Runout flow
30
20
Allowable operating region
10
0
200
400
600
800
1000
1200
1400
1600
0
200
400
600
800
1000
1200
1400
1600
0
200
400
600
800
1000
1200
1400
1600
0
5
0
Power (hp)
60
Design point
80
10
70
76 Best efficiency point
100
40
80
Minimum flow
Efficiency (%)
140
50
25
0
US gpm
FIGURE 2.10
H-Q, BHP, efficiency, and NPSHr curves for a pump with a given speed and impeller diameter.
Hydraulics, Selection, and Curves
35
(H–Q) curve (blue curve in Figure 2.10) typically rises toward shutoff, with
the pump developing lower flows at higher heads, and vice versa. The horsepower curve (red curve in Figure 2.10) typically is rising as flow increases,
although Section VII to follow illustrates that this is not always the case.
Finally, the pump efficiency curve (green curve in Figure 2.10) shows that the
efficiency varies with flow, rising to a peak value known as the best efficiency
point (BEP).
Also shown in Figure 2.10 are other key landmarks that include design
point, shutoff head, runout flow, minimum flow, preferred operating region, and
allowable operating region. These terms will be more thoroughly discussed
in later sections of this book. Figure 2.10 includes the pump NPSHr curve
(brown curve in Figure 2.10). The significance of this curve will be discussed
in Section VI to follow.
Note that where color is shown on pump curves throughout this book, the
color scheme as shown in Figure 2.10 will be used (blue for H–Q curves, red
for BHP curves, green for efficiency curves, and brown for NPSHr curves.
System head curves, which are introduced later in this chapter, will be
shown in purple.
It is only after the performance test described above has been completed
that the manufacturer finally knows whether the performance objectives
for the pump have been reached (i.e., what is the best efficiency flow, head,
horsepower, and pump efficiency, and how these values vary over the full
range of performance of the pump).
The next step for the manufacturer is to perform additional performance
tests with reduced impeller diameters. The pump is disassembled, the impeller is machined to a smaller diameter, and then the test described above is
repeated. Trim increments may be as little as 1/16 in and as much as several
inches, depending on the size of the impeller.
After several complete performance tests at different impeller diameters
have been conducted by the procedure just described, the manufacturer is
able to finally generate the family of curves for the full performance envelope of the pump, which is then published in the manufacturer’s catalog. This
cataloged family of composite curves for a pump over its range of offered
impeller diameters is illustrated in Figure 2.11.
As exemplified in Figure 2.11, some pump manufacturers display the
information on BHP and efficiency on their cataloged curves in a different format from the way Figure 2.10 displays it for a single impeller diameter. In Figure 2.11, the BHP and efficiency data are plotted using iso curves
(lines of constant BHP and constant efficiency). Other manufacturers
simply show a separate efficiency and BHP curve for each impeller trim
offered. Either method of displaying BHP and efficiency as a function of
pump flow and impeller diameter allows the pump selector to determine
the required impeller diameter and motor size for a particular application.
Iso-horsepower lines, when they are used, normally show only the commercially available electric motor sizes.
36
15" Dia.
250
225
9'
60 65
10'
70
75
11'
78
12'
80 81
14"
200
Total head
50 55
Model 3196
Size 6 × 8 – 15 × LT
Imp. dwg. 256–115
Pattern 55436
Eye area 50 sq. in.
14' 16'
82
175
81
80
11"
125
12
100
75
50
50
HP
60
25
75
HP
10
HP
80
70
12"
150
256–116
55437
19'
82
13"
Steel
Feet
NPSHr 8'
0H
60
78
75
15
5H
0H
P
50
40
P
30
P
20
10
Capacity
0
400
0
100
800
200
1200
300
1600
400
2000
2400
500
Meters
Pump Characteristics and Applications
2800
600
3200
700
gpm
0
m3/h
1780 rpm
FIGURE 2.11
Typical manufacturer’s published performance curve family for a centrifugal pump operating
at a fixed speed and with a range of impeller diameters. (Courtesy of Goulds Pumps, Inc., a
subsidiary of ITT Corporation.)
As an example, using the Figure 2.11 pump curve, if the pump design rating
is 1800 gpm and 175 ft, the Figure 2.11 curve shows that the required impeller
diameter would be just under 14 inches. The motor size can then be chosen
using the BHP curves in Figure 2.11. If the pump flow is never expected to
exceed the design flow rate of 1800 gpm, a motor size of 100 HP can be chosen.
Note however that this assumes a specific gravity of 1.0. Remember that if a
liquid other than water is being pumped, the BHP curve must be adjusted up
or down by the specific gravity of the liquid to be pumped.
As Section IX illustrates, in many cases the pump system allows a pump to
operate over a wide range on its H–Q curve. Often, particularly in industrial
applications, a motor size is chosen so that the pump can operate over the
full range of performance at a given diameter, that is, to the end of the curve.
For the example above, this would lead to a selected motor horsepower of
125 HP. This selection of a motor size to allow operation at any point on the
pump curve for a given impeller diameter is known as a nonoverloading motor
selection, and is considered a good selection criterion by most industrial
users. A less conservative approach that is acceptable in many lighter-duty
applications selects a motor size that is adequate for the design point, and
Hydraulics, Selection, and Curves
37
relies on controls or system limitations to keep the pump flow from going
beyond that which would overload the motor.
Another approach allows the motor to make use of its service factor. The
motor service factor is a design margin used in the design of motors, essentially putting more copper in the motor windings to allow the motor to generate more horsepower than the motor is rated for without causing the motor
to run excessively hot. Typical service factors for industrial motors are 1.10,
1.15, or 1.2. A 100-HP motor with a 1.15 service factor is actually capable of
delivering 115 HP without running so hot that the motor insulation would be
harmed or the motor would fail because of excessive heat.
Most conservative industrial users of pumps select motor sizes so that the
motor does not make use of the service factor at all (i.e., the motor is chosen to
be “nonoverloading” over the entire pump performance range, without making use of the service factor.) This is especially recommended if the pump
is to run continuously. This simply means that the motor service factor lets
the motor run cooler than it otherwise would. Many lighter-duty fractional
horsepower motors have quite high service factors (e.g., 1.5), and it is quite
common with residential pumps and other intermittent service or light-duty
commercial and industrial applications for the pump to make use of the
motor service factor at some points of normal operation on the pump curve.
It is recommended, when sizing and selecting centrifugal pumps, to
choose a pump such that the design duty point (head and capacity) is a small
amount to the left of the BEP on the pump curve. The reason for this is that
the vast majority of pumps are oversized, due to the conservatism used by the
pump selector in arriving at estimates for total head in the system. Because
the actual resulting system head is typically less than that predicted by the
engineer at the time of the pump selection process, the pump will tend to
move to the right on its performance curve, to a point where the total head
requirement is less. If the original selection were made to the right of the BEP,
the lower than predicted pump total head would tend to move the operating
point still further to the right on the performance curve. So, it is preferred
to make the initial selection to the left of the BEP, so that, if the actual head
is less, the pump will move closer to its BEP as it moves to the right on the
curve. This is not a hard and fast rule, but the engineer should bear in mind
that if a pump is selected at a point well to the right of the BEP on the performance curve, and if the actual pump total head is less than predicted, this
will allow the pump to move even further to the right, which could lead to
problems with overloading the motor or cavitation in the pump.
Engineers who are sizing pumps often ask what is the maximum amount
away from the best efficiency point on the pump curve that they should
choose a pump to operate. Refer to Figure 2.10, which defines the preferred
operating region and allowable operating region. One rule of thumb puts the
preferred operating region at between 70% and 120% of the BEP flow (which
varies by impeller diameter) for continuous operation. This is a tighter range
around BEP than the allowable operating region, which is generally given as
38
Pump Characteristics and Applications
the range defined by the envelope of performance shown in Figures 2.7 and
2.8. The preferred operating range is even tighter than this for larger, high
energy pumps, and for pumps with higher suction specific speed, which is discussed in Section VII to follow.
The one set of curves on Figure 2.11 not yet discussed are the NPSHr
curves. These are discussed in Section VI below.
VI. NPSH and Cavitation
A. Cavitation and NPSH Defined
As stated in the overview of this chapter, NPSH or net positive suction head
is probably the most misunderstood aspect of pump hydraulics. It is very
important to understand this concept because NPSH problems are among
the most common causes of pump failures, and are often mistakenly blamed
for failures that are completely unrelated.
NPSH must be examined when using centrifugal pumps to predict the possibility of cavitation, a phenomenon that has both hydraulic and sometimes
destructive mechanical effects on pumps. Cavitation, illustrated in Figure
2.12, is a phenomenon that occurs when vapor bubbles form and move along
the vane of an impeller. (What causes the vapor bubbles to form in the first
place is discussed shortly.) As these vapor bubbles move along the impeller vane, the pressure around the bubbles begins to increase. (Figure 1.5 in
Chapter 1 shows that the local pressure increases as the flow moves along
Rotation
Collapsing bubbles
Vapor bubbles
FIGURE 2.12
Cavitation occurs when vapor bubbles form and then subsequently collapse as they move
along the flow path on an impeller.
39
Hydraulics, Selection, and Curves
the path of the impeller vane.) When a point is reached where the pressure
on the outside of the bubble is greater than the pressure inside the bubble,
the bubble collapses. It does not explode, it implodes. This collapsing bubble
is not alone, but is surrounded by hundreds of other bubbles collapsing at
approximately the same point on each impeller vane.
The phenomenon of the formation and subsequent collapse of these vapor
bubbles, known as cavitation, has several effects on a centrifugal pump. First,
the collapsing bubbles make a distinctive noise that has been described as a
cracking or popping or rattling sound, or a sound like the pump is pumping gravel. This can be a nuisance in an extreme situation where a cavitating
pump is operating where people are working. This physical symptom is usually the area of least concern with cavitation, however. Of far greater concern
is the effect of cavitation on the hydraulic performance and the mechanical
integrity of the pump.
The hydraulic effect of a cavitating pump is that the pump performance
drops off of its expected performance curve, referred to as break away, as
illustrated by Figure 2.13, producing a lower than expected head and flow.
An even more serious effect of cavitation is the mechanical damage that
can occur due to excessive vibration in the pump. This vibration is due to
the uneven loading of the impeller as the mixture of vapor and liquid passes
through it, and to the local shock wave that occurs as each bubble collapses.
The shock waves can physically damage the impeller, causing the removal
of material from the surface of the impeller. The amount of material removed
varies, depending on the extent of the cavitation and the impeller material.
If the impeller is made of ferrous-based material such as ductile iron, material is removed from the impeller due to a combination of corrosion of the
ferrous material from the water being pumped and the erosive effect of the
cavitation shock waves. If the impeller material is more corrosion resistant
TH
(ft)
Break
away
Q (gpm)
FIGURE 2.13
Material loss from impeller vane due to cavitation.
40
Pump Characteristics and Applications
but softer, ordinary bronze, for example, the damage that cavitation causes
is similar to a peening operation, in which a piece of relatively soft bronze
is repeatedly struck with a small ball peen hammer. Materials such as 316
stainless steel, with superior corrosion resistance and ability to work harden
under the peening action, have a better ability to resist the metal loss associated with cavitation.
In any case, the removal of material, if it occurs at all, proceeds as long as
the pump is cavitating. Pits can be formed gradually on the impeller vanes
and, in the extreme, the removal of material can actually cause a hole to be
eaten clear through an impeller vane, as Figure 2.14 illustrates. This removal
of material from the impeller has the obvious effect of upsetting the dynamic
balance of the rotating component. The result is similar to what happens if
an automobile tire is not properly dynamically balanced, or if it loses one of
the balance weights, causing excessive vibration.
It is very important to remember that excessive vibration from cavitation
can occur even without the material loss from the impeller described above.
This is true because the vibration from cavitation is caused by the uneven
loading of the impeller and the local shock wave, as mentioned previously,
as well as by the removal of material.
Often, the excessive vibration caused by cavitation subsequently causes
a failure of the pump’s seal and/or bearings. This is the most likely failure
mode of a cavitating pump and the reason why NPSH and cavitation must
be properly understood by the system designer and pump user.
What causes the formation of the vapor bubbles in the first place, without
which the cavitation would not have a chance to occur? To a person who
FIGURE 2.14
Material loss from impeller vane due to cavitation.
Hydraulics, Selection, and Curves
41
has never studied thermodynamics, the most obvious way to create vapor
bubbles—that is, to make a liquid boil—is by raising the temperature of the
liquid. However, this is not what occurs in a cavitating pump because, in the
higher flow range where cavitation is likely to occur, the temperature of a
liquid as it moves through a centrifugal pump remains very nearly constant.
Another way to make a liquid boil, without increasing its temperature, is
if the pressure of the liquid is allowed to decrease. This is due to a thermodynamic property of liquids known as vapor pressure. The vapor pressure
characteristics of water are illustrated in Table 2.3. Note that vapor pressure
can be expressed in psi, or converted to feet of liquid using Equation 2.7. Both
values are shown in Table 2.3.
Every liquid has a characteristic vapor pressure that varies with temperature, as Table 2.3 shows for water. Many handbooks carry these data for various liquids. Figure 2.15 shows the vapor pressure for a number of liquids as
a function of temperature. As described in Chapter 3, Section III, some computer software generated data tables are available that contain vapor pressure data as well as other properties for many liquids.
For any liquid, as temperature goes up, vapor pressure increases. One way
to interpret the vapor pressure data for a liquid is that it shows the temperature at which the liquid boils when it is at a certain pressure. For example,
from Table 2.3, we see that at 14.7 psia (atmospheric pressure at sea level),
water boils at 212°F (100°C). At higher elevations, where atmospheric pressure is less than 14.7 psia, water boils at several degrees below 212°F (100°C).
If water is subjected to a pressure of 90 psia, the liquid does not boil until it
reaches a temperature of 320°F (160°C). This is the principle upon which a
pressure cooker is based. With the pressure cooker operating at a pressure
above atmospheric pressure, the liquid boils at a much higher temperature
than it would in an open pot on the stove, so the food in the pressure cooker
cooks faster.
If a liquid is at a certain temperature in a pressurized container, and the
pressure in the container is allowed to drop below the vapor pressure of the
liquid at that particular temperature, the liquid boils. As an example (using
Table 2.3), if water at 300°F (148.9°C) is in a vessel that is maintained at a pressure of 100 psia, the water is in a liquid state, i.e., is not boiling. However, if
the pressure in the vessel is allowed to drop, when it goes below 67 psia (the
vapor pressure at 300°F), the liquid begins to boil.
The above example is exactly analogous to what can occur in a pump system, causing the creation of vapor bubbles and setting up the conditions for
the pump to cavitate. In a pump system, as the liquid leaves the supply vessel
and approaches the suction of the pump, the local pressure at every point in
the suction line varies, due to changes in elevation and friction in the suction
pipe, valves, filters, and fittings. If this combination of changes in local pressure allows the pressure of the liquid to drop below the vapor pressure at the
pumping temperature, vapor bubbles form, and the conditions are present
for cavitation to commence.
42
Pump Characteristics and Applications
TABLE 2.3
Properties of Water at Various Temperatures
Temperature
(°F)
32
40
45
50
55
60
65
70
75
80
85
90
95
100
110
120
130
140
150
160
170
180
190
200
212
220
240
260
280
300
320
340
360
380
400
420
440
Temperature
(°C)
Specific
Gravity 60°F
Reference
0
4.4
7.2
10.0
12.8
15.6
18.3
21.1
23.9
26.7
29.4
32.2
35.0
37.8
43.3
48.9
54.4
60.0
65.6
71.1
76.7
82.2
87.8
93.3
100.0
104.4
115.6
126.7
137.8
148.9
160.0
171.1
182.2
193.3
204.4
215.6
226.7
1.002
1.001
1.001
1.001
1.000
1.000
0.999
0.999
0.998
0.998
0.997
0.996
0.995
0.994
0.992
0.990
0.987
0.985
0.982
0.979
0.975
0.972
0.968
0.964
0.959
0.956
0.948
0.939
0.929
0.919
0.909
0.898
0.886
0.874
0.860
0.847
0.833
Wt. in lb/
ft3
Vapor
Pressure
(psi
absolute)
Vapor
Pressurea (Feet
Absolute (At
Temperature)
62.42
62.42
62.40
62.38
62.36
62.34
62.31
62.27
62.24
62.19
62.16
62.11
62.06
62.00
61.84
61.73
61.54
61.39
61.20
61.01
60.79
60.57
60.35
60.13
59.81
59.63
59.10
58.51
58.00
57.31
56.66
55.96
55.22
54.47
53.65
52.80
51.92
0.0885
0.1217
0.1475
0.1781
0.2141
0.2563
0.3056
0.3631
0.4298
0.5069
0.5959
0.6982
0.8153
0.9492
1.275
1.692
2.223
2.889
3.718
4.741
5.992
7.510
9.339
11.526
14.696
17.186
24.97
35.43
49.20
67.01
89.66
118.01
153.04
195.77
247.31
308.83
381.59
0.204
0.281
0.340
0.411
0.494
0.591
0.706
0.839
0.994
1.172
1.379
1.617
1.890
2.203
2.965
3.943
5.196
6.766
8.735
11.172
14.178
17.825
22.257
27.584
35.353
41.343
60.77
87.05
122.18
168.22
227.55
303.17
398.49
516.75
663.42
841.17
1056.8
(continued)
43
Hydraulics, Selection, and Curves
TABLE 2.3 (Continued)
Properties of Water at Various Temperatures
Temperature
(°F)
460
480
500
520
540
560
580
600
620
640
660
680
700
705.4
Temperature
(°C)
Specific
Gravity 60°F
Reference
237.8
248.9
260.0
271.1
282.2
293.3
304.4
315.6
326.7
337.8
348.9
360.0
371.1
374.1
0.818
0.802
0.786
0.766
0.747
0.727
0.704
0.679
0.650
0.618
0.577
0.526
0.435
0.319
Wt. in lb/
ft3
Vapor
Pressure
(psi
absolute)
Vapor
Pressurea (Feet
Absolute (At
Temperature)
51.02
50.00
49.02
47.85
46.51
45.3
43.9
42.3
40.5
38.5
36.0
32.8
27.1
19.9
466.9
566.1
690.8
812.4
962.5
1133.1
1325.8
1542.9
1786.6
2059.7
2365.4
2708.1
3093.7
3206.2
1317.8
1628.4
1998.2
2446.7
2972.5
3595.7
4345.0
5242.0
6341.0
7689.0
9458.0
11,878.0
16,407.0
23,187.0
Source: J.H. Keenan: Steam Tables—Thermodynamic Properties of Water, Including Vapor, Liquid,
and Solid Phase. 1969. Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced
with permission.
a Vapor pressure in feet of water (absolute) converted from PSIA using specific gravity at
temperature.
In analyzing a pump operating in a system to determine if cavitation is
likely, there are two aspects of NPSH to consider: NPSHa and NPSHr.
1. NPSHa
Net positive suction head available (NPSHa) is the suction head present at the
pump suction over and above the vapor pressure of the liquid. NPSHa is a
function of the suction system and is independent of the type of pump in the
system. It should be calculated by the engineer or pump user, and supplied
to the pump manufacturer as part of the application criteria or pump specification. The formula for calculating NPSHa is
NPSHa = P ± Hs – Hf – Hvp
(2.19)
P is the absolute pressure on the surface of the liquid in the suction vessel,
expressed in feet of liquid. Hs is the static distance from the surface of the
liquid in the supply vessel to the centerline of the pump impeller, in feet;
the term is positive if the pump has a static suction head, and negative if
the pump has a static suction lift. Hf is the friction loss in the suction line,
44
200
100
80
60
50
40
30
rb
Ca
ne
a
h
Et
e
len
hy
140
100
80
60
50
40
30
20
14
10
5
2
0'' 0
5''
10''
15''
itr
ou
s
ox
i
de
Et
200
N
e
20
e n
en a
yl rop
op P
Pr
on
ia
10
8
6
5
4
3
m
Am
hy
ide
iox
Su
lfu
2
ne
ta
Bu
20''
22.5''
25''
26''
27''
28''
28.5''
Di
ch
lor
1.0
.80
.60
.50
.40
.30
ne
ta
bu
o
Is
e
rid
lo
h
lc
et
M
rd
Absolute pressure (lb/in2)
ide
iox
d
on
.20
.10
–180
–150
–120
–90
–60
–30
Temperature (°F)
0
Gauge pressure (lb/in2)
e
an
eth
M
985
800
600
500
400
300
30
60
Vacuum (in Hg)
1000
800
600
500
400
300
Et
hy
M M
lc
e
hl
oe ethy thy
or
th
lf
l
e
id
yle ne or
e
m
ne ch
a
(tr lor te
an id
s) e
(a)
Pump Characteristics and Applications
29''
29.1''
29.2''
29.3''
29.4''
29.5''
29.6''
29.7''
29.72''
FIGURE 2.15
Vapor pressure of various fluids as a function of temperature. (Courtesy of Gas Processors
Suppliers Association, Tulsa, OK.)
including all piping, valves, fittings, filters, etc., expressed in feet of liquid.
This term varies with flow, so NPSHa must be calculated based on a particular flow rate. Hvp is the vapor pressure of the liquid at the pumping temperature, expressed in feet of liquid absolute.
In a new pump application, NPSHa (and the static term Hs in the above
formula) must be given to the manufacturer with reference to some known
datum point such as the elevation of the pump mounting base. This is
because the location of the pump impeller centerline elevation is generally
not known when the NPSHa calculations are made. It is important that the
datum point of reference be mentioned in the specification, as well as the
calculated value of NPSHa.
45
1000
800
600
500
400
300
e
lo
ch
o
on
M
e
n
pa
o
ia
Pr
y
eth
M
e
rid
r
lfu
Su
e
tan
Bu
l
hy
Et
ne
len
e
200
e
rin
140
100
80
py
Pro
lo
h
lc
e
lfid
su
lo
Ch
ta
bu
Iso
ro
d
Hy
u
rifl
rot
n
ge
60
50
40
30
20
14
10
5
2 0
5''
10''
15''
te
ma
for
yl
eth her
M et
yl
eth
e
id
lor
ch
e
xid
dio
Di
e)
i
an
or
ne
eth
hl
to
m
c
e
o
c
e
ne
A
lor
6)
len
yle
ich
10
hy ne
eth
1.
(tr
et yle
o
=
r
M th
m
°F
e
or
hlo
70
of e
ro
ric
at
lo
lor rid T
R.
Ch hlo
ich
.G
D
c
SP
tra
te
r(
n
te
o
a
w
rb
vy
Ca
ea
H
de
20''
22.5''
25''
26''
27''
S)
10
8
6
5
4
3
rbo
ane ne
Ethetha
m
oro
Am
m
on
Absolute pressure (lb/in2)
20
e
xid
xid ous o
tr
Ni
io
nd
Ca
200
100
80
60
50
40
30
985
800
600
500
400
300
et
hy
len
e(
CI
2
28''
28.5''
Di
W
ch
ate
lo
r
ro
1.0
.80
.60
.50
.40
.30
.20
.10
–60
–30
0
30
60
90
120
150
180
210
240
Vacuum (in Hg)
(b)
Gauge pressure (lb/in2)
Hydraulics, Selection, and Curves
29''
29.1''
29.2''
29.3''
29.4''
29.5''
29.6''
29.7''
29.72''
Temperature (°F)
FIGURE 2.15
(Continued) Vapor pressure of various fluids as a function of temperature. (Courtesy of Gas
Processors Suppliers Association, Tulsa, OK.)
Engineers and pump users are sometimes confused as to how to calculate
NPSHa for a completely closed system such as a chilled water circulation
system, where there is no vented suction tank that establishes a reference
point for measuring the static suction head and the pressure at the surface
of the liquid, the terms Hs and P in Equation 2.19. In general, it is not good
practice to have a completely closed system with no place for the liquid to
expand in the event of temperature fluctuations in the system. Even minor
temperature fluctuations can cause the liquid to expand in the system. If
there is no place for this liquid to go, it can easily cause system pipes to burst.
Good design practice in a completely closed system calls for the introduction
46
Pump Characteristics and Applications
of an expansion tank with a bladder in it with air on one side, to allow some
expansion and contraction of the liquid due to temperature fluctuations. This
not only provides a safety net to keep the pipes from overpressuring due to
the expansion of the liquid, but also provides the reference point needed
to begin the NPSHa calculation. The usual practice is to have the level of
the liquid in the expansion tank be the reference point for the calculation
of the P and Hs terms and the beginning point for the determination of Hf
in the NPSHa calculation. An alternative is to have the tank be a vented tank
located at the high point in the system. The actual location of the expansion
tank is not critical, although usual practice locates it near the suction of the
pump. In some cases, if the liquid being pumped is fairly close to its boiling
point, it may be necessary to actually pressurize the air side of the expansion
tank to increase the overall pressure in the system and provide the liquid
with more margin above its boiling point.
2. NPSHr
Net positive suction head required (NPSHr) is the suction head required at the
impeller centerline over and above the vapor pressure of the liquid. NPSHr
is strictly a function of the pump inlet design, and is independent of the
suction piping system. The pump requires a pressure at the suction flange
greater than the vapor pressure of the liquid because merely getting the liquid to the pump suction flange in a liquid state is not sufficient. The liquid
experiences pressure losses when it first enters the pump, before it gets to the
point on the impeller vane where pressure begins to increase. These losses
are caused by frictional effects as the liquid passes through the pump suction nozzle, moves across the impeller inlet, and changes direction to begin
to flow along the impeller vanes.
NPSHr is established by the manufacturer using a special test, and the
value of NPSHr is shown on the pump curve as a function of pump capacity.
For a pump to operate free of cavitation, NPSHa must be greater than
NPSHr. In determining the acceptability of a particular pump operating in
a particular system with regard to NPSH, the NPSHa for the system must be
calculated by the system designer or pump user, and then the NPSHr for the
pump to be used must be examined at the same flow rate by looking at this
information on the pump curve. This comparison should be made at all possible operation points of the pump, with the worst case usually being at the
maximum expected flow.
Some manufacturers use iso curves to show NPSHr (e.g., Figure 2.11), similar to the treatment of BHP and pump efficiency. Regardless of how it is
shown on the curve, NPSHr increases at higher flow rates, due to the increased
amount of friction loss inside the pump inlet before the liquid reaches the
impeller. For some pumps, NPSHr is also higher with flow unchanged but
impeller diameter reduced, as illustrated on Figure 2.22. This is an impor-
Hydraulics, Selection, and Curves
47
tant point to check, particularly if a reduction in impeller diameter is being
contemplated for an installed pump.
To be conservative, the value of NPSHa should normally be calculated at
its minimum. This means that it should be calculated based on the lowest
liquid level in the supply vessel (minimum static head or maximum static
lift), with the highest friction losses in the suction system (usually at the
highest planned capacity), and at the highest expected liquid temperature
(highest vapor pressure). At the same time, however, one should not be so
conservative in NPSHa calculations as to unnecessarily restrict the type
of pump that can be used. As an example, if a system is being considered
that has large changes in static head, and where the low level would only
incur infrequently and for very short periods of time, it may not be the
best advice to use this low liquid level to calculate the value of NPSHa
given to the pump manufacturer. This is especially the case when doing
so might force a different pump than would otherwise be preferred. This
could result in a pump selection that might cost significantly more, use
more energy, or require more maintenance. This points out the importance
of good communication between the pump user and supplier when the
pump selection is made.
B. Calculating NPSHa: Examples
Several examples illustrate how to calculate NPSHa, how to determine if a
pump might experience cavitation, and what to do about it if it does. Also
refer to the discussion in Chapter 3, Section III to learn how computer software can be used to perform these calculations.
Example 2.2: Calculating NPSHa and Comparing with Published NPSHr
PROBLEM
Calculate NPSHa for this system and verify the adequacy of the selected
pump.
GIVEN
The system shown in Figure 2.2
Pump shown in Figure 2.11
Static suction lift = 12 ft
Design capacity (Q) = 2000 gpm
Design pump total head = 175 ft
Liquid = water at 80°F (SG = 1.0)
Hf = 3 ft (Usually, this is calculated using friction tables as illustrated in Example 2.1 in Section IV of this chapter.)
P = 14.2 psia (atmospheric pressure at the pump site)
48
Pump Characteristics and Applications
SOLUTION
NPSHa = P ± Hs – Hf – Hvp (Equation 2.19)
P = 14.2 psia (given)
= (14.2 × 2.31)/1.0 ft
= 32.8 ft
Hs = –12 ft (given; negative because it is a suction lift)
Hf = 3 ft (given)
Hvp = 1.2 ft (Table 2.3 at 80°F)
NPSHa = 32.8 – 12 – 3 – 1.2 = 16.6 ft
From Figure 2.11, at 2000 gpm, NPSHr = 11.2 ft, so NPSHa > NPSHr, and the
chosen pump is acceptable from an NPSH point of view (i.e., does not cavitate).
Example 2.3: Calculating NPSHa and Comparing with Published NPSHr
PROBLEM
Calculate NPSHa for this system and verify the adequacy of the selected
pump.
GIVEN
The problem presented in Example 2.2 above, except using water at 160°F
SOLUTION
NPSHa = P ± Hs – Hf – Hvp (Equation 2.19)
SG = 0.98 (Table 2.3 at 160°F)
P = 14.2 psia (given)
= (14.2 × 2.31)/0.98 ft
= 33.5 ft
Hs = –12 ft (given; negative because it is a suction lift)
Hf = 3 ft (given)
Hvp = 11.2 ft (Table 2.3 at 160°F)
NPSHa = 33.5 – 12 – 3 – 11.2 = 7.3 ft
From Figure 2.11, at 2000 gpm, NPSHr = 11.2 ft, so NPSHa < NPSHr, and the
chosen pump is unacceptable from an NPSH point of view (i.e., the pump
would cavitate). In fact, most pumps in the flow range of 2000 gpm require
more than 7.3 ft of NPSH, so it would be difficult to locate a standard commercially available pump for this service as described.
C. Remedies for Cavitation
To make the proposed pump in Example 2.3 above acceptable from an NPSH
point of view, there are a number of system modifications that might be
Hydraulics, Selection, and Curves
49
considered to increase the calculated NPSHa. These alternatives might include
system changes that affect any of the four terms of Equation 2.19, or some combination of them. Usually all of the system changes to be considered cost money,
and often the engineer must compare the overall costs of different alternatives.
The most obvious system change in Example 2.3 above would be to change
the value of Hs to a smaller negative number by raising the suction supply
vessel, or the minimum liquid level in the suction supply vessel, or by moving the pump to a lower elevation.
A second system change might involve reducing the value of Hf by making the suction piping shorter (moving the pump closer to the supply vessel).
Alternative system modifications could involve making the suction pipe size
larger, eliminating fittings or valves in the suction line, or, if there is a filter
in the suction line, changing the filter design to a type that does not permit
as high a differential pressure buildup between cleaning operations. In general, filters are better located on the discharge side of the pump for this reason, unless they are specifically installed to keep solids of a size that could
damage the pump from entering the pump.
A third system change might involve reducing the temperature of the process liquid (because it is the higher temperature that causes the pump in
Example 2.3 to be unacceptable from an NPSH point of view).
Finally, the first term in Equation 2.19 could be increased by pressurizing
the tank with a blanket of air, nitrogen, or other compatible gas above the
liquid surface. The gas blanket could be of fairly low pressure (only several
pounds of pressure).
However, a blanket gas as a substitute for a static head tank in a pump
system should be used with caution. If some of the blanket gas is permitted to dissolve in the liquid (a function of the gas solubility), the dissolved
gas can be liberated in the low pressure area at the entrance of the pump,
and this can reduce the effective NPSHa. The amount of additional margin
between NPSHa and NPSHr to account for the release of this dissolved gas
can be substantial (Ref. [7]). Note that to the author’s knowledge, none of the
software programs discussed in Chapter 3 for modeling piping systems take
into account the possible effects of dissolved gas on the NPSHa calculation.
If no combination of the changes to the system suggested in the preceding paragraphs can be done, or if they are not adequate to raise the value
of NPSHa to a level where it exceeds NPSHr, there may be other options to
consider. Recall from the definitions of NPSHa and NPSHr, which are given
prior to the examples above, that both NPSHa for a system and NPSHr for
a pump vary with flow, with NPSHa decreasing and NPSHr increasing at
higher flows. This relationship is shown in Figure 2.16. At flows higher than
the intersection point of the two curves in Figure 2.16, the NPSHr would
exceed the NPSHa and so the pump in that system would cavitate. The system modifications discussed above to raise NPSHa would have the effect of
shifting the NPSHa curve in Figure 2.16 up, which would move to a higher
flow the point at which cavitation would be present. Figure 2.16 suggests
50
Pump Characteristics and Applications
No cavitation
(NPSHa > NPSHr)
Cavitation
(NPSHr > NPSHa)
NPSH
(ft)
Required
Available
H
(ft)
Q (gpm)
FIGURE 2.16
Variation of NPSHa and NPSHr with flow through the system.
one rather rudimentary way to solve the problem of a pump in the field that
is cavitating, or the problem found in Example 2.3 above where a proposed
pump is unacceptable from an NPSH point of view. If a lower flow rate is
acceptable from a process point of view, it may be possible to merely throttle
the pump back to a lower flow rate by partially closing a valve in the pump
discharge, until the flow is reduced to the point where NPSHr is reduced and
NPSHa is increased enough that the pump no longer cavitates. Naturally an
appropriate type of throttling valve needs to be used to do this.
If none of the above proposed solutions are adequate to solve the problem,
then the only way to eliminate the cavitation problem described here is with
a different pump. There may be a commercially available pump that has a
lower NPSHr. If a pump is available to make the same head and flow but at a
slower speed, the slower speed pump would most likely have a lower NPSHr
(although it would also be a physically larger pump, and would likely be
more expensive). Another possibility that may be available without going to
a slower speed is a double suction pump. (Refer to Chapter 4, Section II.B.)
There are also specially designed low NPSHr, single suction pumps available from certain manufacturers in limited styles and sizes of process pumps.
These pumps achieve the lower value of NPSHr by having a special impeller
design with an increased inlet (eye) area. Although this solution may be the
most economical choice, or perhaps the only alternative in certain circumstances, care should be taken when using these special impeller designs. The
larger inlet area can restrict the flow range over which the pump can operate
successfully to a smaller range than a comparable pump with a “standard”sized impeller inlet. The geometry of the impeller inlet is described by a
Hydraulics, Selection, and Curves
51
dimensionless variable called suction specific speed. This subject is discussed
again in Section VII to follow and in Chapter 4, Section II.C.
It may be necessary to divide the flow between two pumps operating in
parallel to solve the problem raised by this example. This would be especially worth considering if there is a range of flow requirement. Or, it may
be possible to add a second pump in series upstream of the first one, acting
as a booster pump to increase the NPSHa at the inlet of the second pump.
Applications involving multiple pumps in parallel or series are discussed
further in Sections X and XI to follow.
If all possible affordable alternatives have been explored for modifying
the system and for changing the pump, and still no solution has been found,
in some cases there is no choice but to live with the cavitation. For example,
if the expected flow, temperature, friction losses, and liquid levels are such
that it is estimated that the pump will cavitate during only 5% to 10% of the
time it operates, the designer or user may simply choose to live with this.
Also, it may be possible to minimize the damage from the cavitation by
selecting an impeller material that is more resistant to cavitation damage,
such as a high chrome stainless steel, and a pump with a more robust bearing system.
Also, in considering the possibility of designing a system with a pump
which may be operating in a cavitating mode for certain periods of time,
recall that part of the cavitation damage occurs as a result of the shock wave
caused when the vapor bubbles collapse. It turns out that the amplitude of
the shock wave produced in a cavitating pump is directly proportional to
the value of the pump NPSHr. Thus, a pump that requires only 6 ft of NPSH
would cause much less damage when cavitating than would a cavitating
pump of the same metallurgy that requires 25 ft of NPSH.
D. More NPSHa Examples
Several more examples help to further solidify the reader’s understanding
of NPSH.
Example 2.4: Calculation of NPSHa and Suction Pressure
PROBLEM
Calculate NPSHa for this system.
GIVEN
The system shown in Figure 2.17
Suction head = 7 ft
Liquid = water at 300°F
Hf = 2 ft
P = 75 psia
Atmospheric pressure = 14.7 psia
52
Pump Characteristics and Applications
P
HS
CL
FIGURE 2.17
System for example problem on calculating NPSHa and suction pressure (Example 2.4).
SOLUTION
NPSHa = P ± Hs – Hf – Hvp (Equation 2.19)
SG = 0.92 (Table 2.3 at 300°F)
P = 75 psia (given)
= (75 × 2.31)/0.92 ft
= 188.3 ft
Hs = 7 ft (given; positive because it is a suction head)
Hf = 2 ft (given)
Hvp = 168.2 ft (Table 2.3 at 300°F)
NPSHa = 188.3 + 7 – 2 – 168.2 = 25.1 ft
If the pressure in the supply vessel in the proceeding example were
changed to 67 psia instead of 75 psia, the first term in the NPSHa calculation
would be 67 × 2.31/0.92 = 168.2 ft. This would exactly cancel out the fourth
term, Hvp, because 67 psia is also the vapor pressure of water at the pumping
temperature of 300°F. In that case,
NPSHa = Hs – Hf = 7 – 2 = 5 ft
In the case of the example above, the pressure in the supply vessel could
get no lower than 67 psia because if it did, the vessel contents would no longer be in a liquid state. The situation just considered, where the pressure in
the supply vessel was changed to 67 psia, illustrates an application involving pumping a saturated liquid. This is a common application in industry, for
example, a condensate pump in a power plant. When a saturated liquid is
to be pumped, the first and fourth terms of the NPSHa formula (P and Hvp)
cancel out each other, so the NPSHa is only equal to:
NPSHa = Hs – Hf (for saturated liquids)
(2.20)
Hydraulics, Selection, and Curves
53
Because Hf is always subtracted, the only term that adds positively to NPSHa
is Hs, which is why the condensate pump must be located at the lowest possible elevation in the plant to get the value of Hs as large as possible.
Some people mistakenly believe that as long as a pump has a reasonable
amount of suction pressure (regardless of temperature), then cavitation is
not a concern. This is not true, as an example illustrates. Suppose we wish
to calculate what suction pressure would be measured by a gauge located at
the centerline of the suction of the pump in Example 2.4 above, but operating
with a supply vessel pressure of 67 psia. Ignoring velocity head, which is a
very small term, the expression for suction pressure is
Suction pressure = P + Hs – Hf
(2.21)
For P = 67 psia (168.2 ft), Hs = 7 ft, and Hf = 2 ft
Suction pressure = 168.2 + 7 – 2 = 173.2 ft
Converting to psi, 173.2 × 0.92/2.31 = 69.0 psia = 54.3 psig
Although suction pressure is shown above to be well over 50 psig, NPSHa
was calculated in Example 2.4 as only 5 ft, so cavitation may be a problem
with this system. The important point here is that one should not confuse
the existence of seemingly adequate suction pressure with the question
of whether or not a pump cavitates in a given service. Cavitation does not
depend on suction pressure, but rather on the relative size of NPSHa for a
system and NPSHr for a pump operating in that system.
Example 2.5: Calculation of NPSHa with Suction Source under Vacuum
PROBLEM
Calculate NPSHa for this system.
GIVEN
The system shown in Figure 2.9
Suction head = 10 ft
Liquid = water at 140°F
Hf = 1.4 ft
Vacuum in supply tank = 5 in Hg
Atm. press. = 14.5 psia = 29.5 in Hg
SOLUTION
NPSHa = P ± Hs – Hf – Hvp (Equation 2.19)
54
Pump Characteristics and Applications
SG = 0.985 (Table 2.3 at 140°F)
P = (29.5 – 5) × 1.133/0.985 = 28.2 ft (Equation 2.8)
Hs = 10 ft (given; positive because it is a suction head)
Hf = 1.4 ft (given)
Hvp = 6.8 ft (Table 2.3 at 140°F)
NPSHa = 28.2 + 10 – 1.4 – 6.8 = 30.0 ft
E. Safe Margin NPSHa versus NPSHr
An often-asked question is: “What is a safe margin to maintain between
NPSHa and NPSHr?” Unfortunately, like so many questions related to pumps,
the answer must begin with “That depends.…” Section VI.C above covers the
fact that for pumps with very low NPSHr values and with impellers constructed of cavitation resistant materials, it is possible to operate under some
amount of cavitation for an extended period of time without causing damage
to the pump impeller (although the other symptoms of cavitation—noise,
vibration, and the drop off of the H–Q curve from its expected path—would
still be present). This situation is the exception rather than the rule, however.
For the majority of applications, it is good practice to have a safe margin
between the available and required NPSH.
Fortunately, most pump applications have a safe margin, with an NPSHa of
at least 10 ft above that required by most commercially available pumps. In
most circumstances, it is usually only applications with excessive amounts
of suction line friction losses, high temperature liquids, large amounts of
suction lift, or high flow applications (above 15,000 gpm) that present NPSH
problems.
The way that the NPSHr curve is developed by the manufacturer involves
a special test, done in a different laboratory test setup from the pump performance test described in Section V. The NPSH test is usually done in a closed
loop piping system where the manufacturer has the capability not only to
measure head and flow but also to vary the NPSHa. Varying the NPSHa is
usually accomplished by means of pulling a vacuum at the top of the supply tank (reducing the value of P in the NPSHa equation) while maintaining a constant flow and continuously monitoring head. When the NPSHa
is reduced to the point that the pump begins to cavitate, the pump curve
begins to drop off, as described earlier and shown in Figure 2.13. Because
flow is being held constant during the test, this means that the pump TH
begins to drop.
According to the convention established in the Hydraulic Institute (HI)
Standards (Ref. [4]), when the head drops by 3%, the pump is presumed to
be in full cavitation. Three percent was chosen by the writers of the HI Test
Code because when cavitation first begins, the head fluctuates up and down
and is difficult to measure. There is also the accuracy of the test instrumentation to be considered. This figure (3%) was chosen by the HI as an amount of
head reduction that would unmistakably mean that the pump is cavitating.
Hydraulics, Selection, and Curves
55
When the head drops by 3%, the pump is defined to be in full cavitation and
the NPSHa in the laboratory setup at that flow point is set equal to NPSHr.
This establishes one point on the NPSHr curve for that pump.
The pump control valve is then repositioned and the test is repeated at
a different flow rate. After several test points are taken, the NPSHr curve
can be drawn for the full impeller diameter. Normally this NPSH test is
repeated for each impeller trim when the manufacturer is developing the
catalog curve.
If a pump’s NPSHr curve is developed by the test described above, and if
the pump is operated in a system where NPSHa exactly equals NPSHr, the
pump would actually be cavitating and the performance curve would drop
off, with head dropping by 3% from what it should be. This emphasizes the
need for a margin between NPSHa and NPSHr.
Note that for some pumps, the NPSHr is higher at a given capacity with a
trimmed impeller than it is with a full diameter impeller (see Figure 2.22).
Users should be aware of this if, for example, they choose to trim an impeller in the field to lower the head produced by the pump at a given flow. The
NPSHr may actually be higher at the same flow rate with the trimmed impeller than it was with a full diameter impeller.
The effect of impeller trim on NPSHr can be predicted by studying the
manufacturer’s family of curves for the pump in question. If the NPSHr
curves are presented as iso curves, they are completely vertical if there is no
effect on NPSHr from trimming the impeller (as in the case of Figure 2.11).
They curve to the left for trimmed impeller diameters if the pump is one that
has higher NPSHr at trimmed impeller diameters (as in the case of Figure
2.22). Unfortunately, not all pump manufacturers consistently show on their
curves the increased NPSHr with reduced impeller diameter.
Another consideration when deciding the margin that should be maintained between NPSHa and NPSHr is that testing and experience have shown
that damage due to cavitation when pumping cold water is more severe than
it is when pumping hot water and certain other liquids. Figure 2.18 shows
an NPSH Reduction Chart, which allows the determination of a correction
factor to reduce cataloged NPSHr (or increase calculated NPSHa) under certain conditions. If the liquid is one of the liquids shown in this chart, entering the temperature or vapor pressure permits determination of a correction
factor that effectively allows a closer margin between NPSHa and NPSHr.
However, users of this chart are cautioned that there are a number of limitations that apply to the use of the chart.
No NPSH reduction should exceed the lesser of 50% of the NPSHr with
cold water, or 10 ft. Reductions are not applicable or may be different if there
is entrained air or other noncondensable gas in the liquid, or if the system
has transient changes in temperature or pressure. Vapor pressures of hydrocarbon mixtures should be determined by the bubble point method at pumping temperature, rather than using the Reid vapor pressure or the vapor
pressure of the lightest fraction. Finally, use of the chart for liquids other
56
Pump Characteristics and Applications
1000
600
400
ne
pa
10
200
8
100
6
5
4
3
60
2
40
1.5
20
b
Iso
1.0
ne
a
ut
ne
ta
Bu
0.5
11
ge
ra
nt
R-
10
8
l
Re
fri
6
ho
Vapor pressure (psia)
80
Wa
Me
ter
thy
l al
co
4
2
1
NPSHr reduction (ft)
o
Pr
0
50
100
150
Temperature (°F)
200
300
400
FIGURE 2.18
NPSH reduction chart. (Courtesy of the Hydraulic Institute, Parsippany, NJ; www.pumps.org.)
than those specifically shown on the chart must be considered experimental
only (see Ref. [3]).
Because of all the restrictions and hedges given for the use of this chart,
many users and engineers have simply decided to not take advantage of
the additional correction to NPSH that the chart allows. The corrections
should not be made in any event without consulting with the pump manufacturer, who may have more NPSH test data for that particular pump.
Not using the correction factors allowed by the NPSH reduction chart is a
Hydraulics, Selection, and Curves
57
more conservative approach and should add an additional margin of safety
between NPSHa and NPSHr for some liquids.
Mention has already been made of the fact that if there are dissolved gases
in the liquid, then depending on the solubility of the gases, there may be
a substantial additional margin required between NPSHa and NPSHr to
account for the release of the dissolved gases in the low pressure area at the
impeller inlet (Ref. [7]).
Another point to consider is how oversized is the pump’s bearing system
(bearing load capability and shaft diameter) for the application. A very conservative bearing system for a given application could tolerate greater loads
on the impeller from cavitation without deflecting the shaft at the mechanical seal faces than a less conservative bearing system design. As an example,
a typical ANSI process pump line (see Chapter 4, Section XIV.B) might have
20 to 25 sizes of pumps, but only use three designs of bearing frames for the
entire line. Thus, each bearing frame is used for a range of approximately
eight pump sizes. When a bearing frame is being used on the smallest of
the eight sizes that it can accommodate, its design is much more conservative than if it is used for the largest of the eight sizes. So, when it is used for
the smallest of the eight sizes, it should be able to tolerate relatively greater
impeller loads due to cavitation without causing vibration or deflection of
the shaft at the mechanical seal.
In summary, when considering the margin that should be maintained
between NPSHa and NPSHr for a particular application, the questions to ask
include:
• How conservatively was NPSHa calculated for the system, and for
what percentage of the pump’s duty cycle is this low value of NPSHa
actually present?
• What is the level of NPSH and thus the relative amplitude of the
cavitation shock waves?
• What is the pump impeller material, and how resistant is it to cavitation damage?
• Is the liquid pumped one of those that allows a reduction of NPSHr
by the chart in Figure 2.18?
• How oversized is the bearing system for the application?
• Does the pump system make use of a gas blanket that may become
dissolved in the liquid and subsequently liberated in the low pressure area of the impeller inlet?
Depending on the answers to these questions, the recommended minimum margin between calculated NPSHa and NPSHr can range from 0% to
35%. A conservative rule of thumb is that NPSHa should exceed NPSHr by a
minimum of 5 ft or be equal to 1.35 times the NPSHr, whichever is the greater
value. For example, for an NPSHr of 10 ft, NPSHa should be a minimum of
58
Pump Characteristics and Applications
15 ft. For an NPSHr of 20 ft, NPSHa should be a minimum of 27 ft. As pointed
out at the beginning of this discussion, good engineering practice is to have
a safe margin at all times if possible, and to add more if it can be easily and
economically done. As to the acceptable margin for any particular application, the material in this chapter should give the engineer and pump user the
tools to help make informed decisions in this regard.
F. NPSH for Reciprocating Pumps
Before leaving the subject of NPSH, a final point should be made regarding
an additional factor to be calculated when a reciprocating positive displacement pump is being considered. This type of pump is also subject to cavitation damage (as are all positive displacement pumps) and thus a calculation
of NPSHa must be made and supplied to the pump manufacturer. With this
type of pump, there is an additional term in the formula for NPSHa, called
acceleration head. Acceleration head is caused by the rapid deceleration and
acceleration of the liquid in the suction line as the suction check valves open
and close in the pump. The calculated value of acceleration head must be
subtracted from the calculated NPSHa when a reciprocating pump is being
considered. Acceleration head Ha (in feet) is computed for a given pump and
system combination as follows (Ref. [3]):
Ha =
L × V × rpm × C
g×K
(2.22)
where L is the length of the suction pipe, ft; V is the velocity in the suction
pipe, ft/s; rpm is the pump rpm; C is the factor for number of pistons or plungers; g is the acceleration of gravity, 32 ft/s2; and k is the factor for liquid type.
No. Pistons/Plungers
2
3
4
5
6
7
C
0.115
0.066
0.080
0.040
0.055
0.028
Liquid
K
Water
Petroleum
Liquid with entrained gas
1.4
2.5
1.0
The acceleration head for reciprocating pump systems can often be as
large a factor as all of the other terms in the NPSHa formula combined, so
59
Hydraulics, Selection, and Curves
it should by no means be ignored. If the analysis of a proposed reciprocating pump and system indicates that NPSH may be a problem, possible
solutions might include reducing the length or increasing the diameter of
the suction line, slowing down the pump, or installing a pulsation dampener in the suction line. Pulsation dampeners were described in Chapter
1, Section VI.C.12, to reduce discharge pulsations caused by reciprocating
pumps; but in this case they are referred to as suction stabilizers, and their
function is to help absorb the acceleration and deceleration of the liquid in
the suction line.
VII. Specific Speed and Suction Specific Speed
Specific speed (Ns) is a design index primarily used by pump engineers to
describe the geometry of pump impellers and to classify them as to their
type. It is referred to as a “dimensionless” index, but the term is used loosely,
as described below. An understanding of how to calculate and interpret the
specific speed for a particular pump provides greater insight into the reasons why pump impellers are shaped so differently, why different impellers have such a range of flow and head capability, why impellers have such
differently shaped performance curves, and why there is such wide variation in the value of efficiency at the BEP for different pumps. Furthermore,
Chapter 6, Section II, shows how it is possible to use specific speed to select
pumps for maximum efficiency.
The formula for pump specific speed Ns, in USCS units, is
Ns =
N× Q
H 3/4
(2.23)
where N is the pump speed, rpm; Q is the capacity at BEP, full diameter,
gpm; and H is the pump head per stage at BEP, full diameter, ft.
In the U.S. pump industry, the definition of Q in Equation 2.23 when it
comes to double suction impellers (covered in Chapter 4, Section II.B) is not
treated consistently. For the majority of U.S. pump designers, Q is taken as
the full pump capacity. However, a few designers have employed an alternative calculation using one half of the capacity, and U.S. pump standards do
allow for this alternative calculation When working with manufacturers of
double suction pump impellers, it’s best to clarify this point if specific speed
is being discussed.
Note that the above inconsistency in defining Q does not exist in defining
Q when It comes to calculating suction specific speed (discussed later in this
section), where the value of Q in Equation 2.24 for a double suction impeller
60
Pump Characteristics and Applications
is one half the total pump capacity. European practice uses one half of total
pump capacity for both terms with a double suction impeller.
An analysis of the units in Equation 2.23 reveals that the term Ns is not truly
dimensionless, although it would become so with the addition of g, the acceleration of gravity, into the equation’s denominator (taken to the 3/4 power), and
with appropriate conversion of the terms of the equation into other equivalent
terms. The convention in the centrifugal pump industry is to omit the g term
but still treat Ns as dimensionless. However, this creates the unfortunate consequence of having a different value for specific speed if SI units are employed
in the calculation instead of USCS units. In SI units, the specific speed is designated Nsm and is usually based on capacity expressed in cubic meters per hour
and head in meters. Therefore, Ns = 0.8609 Nsm. If capacity is expressed in cubic
meters per second and head in meters, then Ns = 51.65 Nsm.
The specific speed of a particular pump can be calculated from the pump
curve; picking N, Q, and H off the curve at full diameter, best efficiency
point; and applying Equation 2.23. Once Ns for a particular pump has been
calculated, its value will not change, even if the pump is run at a different
speed. Obviously, if the pump is run at a different speed, the pump’s total
head and capacity do change but the specific speed does not change, because
it is defined by the equation above. In fact, it is the fact that the specific
speed will not change that is the basis for the derivation of the pump affinity
laws (Section VIII), which allow the pump performance to be predicted for
changes in pump speed or impeller diameter.
Table 2.4 shows the calculated value of specific speed based on Equation
2.23 for some arbitrarily selected pump BEP, full diameter conditions. The
data are arranged in order of increasing Ns, the last column in the table. The
data are not, however, arranged strictly in order of increasing Q or decreasing H (although that is the general trend). This is because the formula for Ns
depends on all three variables, rather than on any one variable. For example,
the fourth entry in Table 2.4 has the smallest Q value of all the entries. Yet
because the value of H is so small for that pump, the value of Ns is higher
than the others above it in the table.
Figure 2.19 illustrates typical impeller profiles for the range of specific
speeds found for rotodynamic pumps. At the low end of the range, impellers
TABLE 2.4
Specific Speed Ns for Selected Pumps
Q
H
N
Ns
120
350
1000
70
9000
50,000
300
300
100
20
40
20
3550
3550
1780
3550
1180
590
540
1250
1780
3140
7040
13,960
US units
Metric
Vanes
900
800
700
600
Vanes
Francis-vane area
Mixed-flow area
Vanes
Values of specific speeds
FIGURE 2.19
Impeller profile vs. Ns. (Courtesy of the Hydraulic Institute, Parsippany, NJ; www.pumps.org.)
Vanes
2000
40
Radial-vane area
1500
Hub
3000
60
Hub
4000
80
Hub
6000
Hub
1000
20
Impeller
shrouds
5000
100
500
10
Vanes
7000
Impeller
shrouds
Axial-flow area
US units
Axis of
rotation
Impeller
hub
Metric
8000
9000
10,000
200
Impeller shrouds
15,000
300
150
20,000
400
Impeller shrouds
Hydraulics, Selection, and Curves
61
62
Pump Characteristics and Applications
Axial flow head
Mixed flow head
Radial flow head
BHP
Total head
develop head by moving the liquid radially from the shaft centerline. These
low specific speed pumps are called radial flow pumps. Radial flow pumps
have the characteristic of relatively low flow and high head. At the opposite
end of Figure 2.19, impellers develop head through axial forces, and so these
high specific speed pumps are referred to as axial flow pumps, or propeller
pumps. Axial flow pumps have the characteristic of relatively high flow and
low head. As Ns increases, the ratio of the impeller outlet diameter to inlet
diameter decreases. This ratio becomes 1.0 for a true axial flow impeller.
Pumps that are neither pure radial flow nor pure axial flow are called mixed
flow, and represent some combination of radial flow and axial flow. Actually,
because the total specific speed range for centrifugal pumps is a continuous spectrum from several hundred to about 20,000, pumps with Ns ranging from several hundred to about 2000 are often referred to as radial flow,
pumps with Ns greater than about 8000 are called axial flow, and pumps
with Ns between these two ranges are called mixed flow.
One of the characteristics of specific speed is its effect on the shape and
slope of the pump head–capacity and BHP curves. This is illustrated in
Figure 2.20. Radial flow pumps have the flattest H–Q curves, with the head at
zero flow (called the shutoff head) often no more than about 120% of the head
at BEP. The lower the specific speed, the flatter the pump H–Q curve. Note
that the slope of the H–Q curve is also affected by certain design parameters
in the impeller design such as the number of vanes and the vane angles.
Low specific speed pumps sometimes exhibit a drooping characteristic
at shutoff, as illustrated in Figure 2.21. This may lead to unstable operation
in certain systems, although it may be perfectly acceptable for use in other
circumstances. The dashed lines shown as A and B in Figure 2.21, and the
conditions that might lead to instability with a drooping H–Q curve such
as shown in this figure, are discussed in Section IX on system head curves.
Flow rate
FIGURE 2.20
Slope of H–Q and BHP curves varies with specific speed.
Axial flow BHP
Mixed flow BHP
Radial flow BHP
63
Hydraulics, Selection, and Curves
1
H
(ft)
2
A
3
B
Q (gpm)
FIGURE 2.21
Drooping H–Q curve.
Mixed flow pumps have steeper H–Q curves than radial flow pumps, as
illustrated in Figure 2.20. The head at shutoff for mixed flow pumps is on the
order of 160% of the head at BEP. Axial flow pumps have the steepest H–Q
curves of all, with shutoff head being in the range of 300% of head at BEP.
Another characteristic of some mixed flow pumps is a dip in the H–Q curve
(Figure 2.22). This may or may not be a problem, depending on the shape of
the system head curve (see Section IX). Some manufacturers simply do not
show the dip on their published curves, but stop the H–Q curve short of
going back to zero flow, with a notation that the pump should not be run in
the unstable region.
The BHP curve shape is also affected by specific speed, as shown in
Figure 2.20. Radial flow pumps exhibit increasing horsepower with increasing pump flow, with the maximum BHP occurring at the maximum flow
at which the pump can operate (called runout). Because most process and
transfer pumps are in the radial flow specific speed range, this is the shape
of the horsepower curve with which the majority of pump engineers and
users are most familiar. Mixed flow pumps have a flatter horsepower curve,
and axial flow pumps have their horsepower curve shaped just the opposite
of radial flow pumps, with the highest horsepower at the lowest flow. The
BHP at shutoff for an axial flow pump is in the range of twice the BHP at
BEP. Axial flow pumps are generally not run at low flows, in part because
of this higher horsepower at lower flows. Furthermore, if it has a BHP curve
that rises toward shutoff, the pump is started with the pump discharge valve
open rather than closed or nearly closed, which is the usual valve position
when the pump is started. In most cases, the motor for axial flow pumps is
not sized to handle the higher horsepower at lower flows. Attempting to start
the pump with the valve closed would cause the motor to overload.
64
221/2" × 19 3/4"
110
100
213/8" × 18 5/8"
90
201/4" × 16 1/2"
Total head
80
49
57
65
72
Model
Size
Imp. dwg.
Pattern
Eye area
79 83
85
NPSHR
87
88
3175
18 × 18 – 22 H
D02467A02
63479
233.7
5 vane
19'
88
197/8" × 13 1/4"
60
87
85
83
79
30
72
“For ratings to left,
consult factory”
20
12
10
0
15
5H
17
0H
P
P
20
5H
25
0H
P
2000
4000
0
500
1000
6000
8000
1500
10,000
2000
12,000
2500
14,000
3000
30
0H
0H
P
P
15
10
5
16,000
3500
20
P
Capacity
0
25
25'
50
40
35
30
18'
201/4" × 14 1/2"
70
D02467A01
63075
Meters
120
Steel
Feet
Pump Characteristics and Applications
0
18,000 gpm
4000 m3/h
885 rpm
FIGURE 2.22
Dip in H–Q curve. (Courtesy of Goulds Pumps, Inc., a subsidiary of ITT Corporation.)
Suction specific speed is a design index used to describe the geometry of the
suction side of an impeller, or its NPSHr characteristics. The term “suction
specific speed” is designated Nss or S, and its formula, very similar to the
formula for specific speed, is
S =
N × Q
NPSH 3r /4
(2.24)
As with specific speed, the terms in the equation above are taken at BEP, full
diameter. The value of Q in Equation 2.24 for a double suction impeller (see
Chapter 4, Section II.B) is taken as one half the total pump capacity.
Typical values of S for most standard designed impellers are in the range
of 8000 to 9000. Although there are special impeller designs available from
some manufacturers having higher S values (Section VI.C and Chapter 4,
Section II.C), various sources recommend that, except in special circumstances, pumps be chose with S values below a maximum value. A maximum value of S that is often specified is 10,000. Using higher values of S as
an impeller design criteria will result in lower NPSHr, but could introduce
problems of recirculation (discussed in Chapter 8, Section III.B.5) if the pump
operates at flows less than the BEP flow.
65
Hydraulics, Selection, and Curves
VIII. Affinity Laws
The pump affinity laws are rules that govern the performance of a centrifugal pump when the speed or impeller diameter is changed. The basis for
the derivation of the affinity laws is that a pump’s specific speed, once calculated, does not change. If the performance of a pump at one speed and
impeller diameter are known, it is possible to predict the performance of the
same pump if the pump’s speed or impeller diameter is changed.
There are two sets of affinity laws. With the impeller diameter, D, held
constant, the first set of laws is as follows:
Q1
N
= 1
Q2 N2
H1  N1 
=
H 2  N 2 
(2.25)
2
(2.26)
BHP1  N 1 
=
BHP2  N 2 
3
(2.27)
With the speed, N, held constant, the second set of laws is
Q1 D1
=
Q2 D2
H1  D1 
=
H 2  D 2 
(2.28)
2
BHP1  D1 
=
BHP2  D 2 
(2.29)
3
(2.30)
where Q is the capacity, H is the total head, BHP is the brake horsepower, N
is the pump speed, and D is the impeller outside diameter.
The manufacturer or user may make use of the second set of affinity laws
above to calculate the exact impeller trim on a pump to make a particular rating, if the performance at specific impeller diameters bracketing the rating is
66
Pump Characteristics and Applications
known, or if the performance at only one impeller diameter is known and performance at another diameter is to be determined. Realistically, for most applications, the required impeller trim can be “eyeballed” by interpolating between
the diameters shown on the family of tested H–Q curves for the pump.
Note that the use of the second set of affinity laws, Equations 2.28, 2.29, and
2.30, to calculate impeller trims generally produces results that dictate more of
an impeller trim than is actually required to produce the desired head and flow
reduction, with the amount of error being as high as 15% to 20% in some cases.
Lower specific speed pumps generally tolerate a greater amount of impeller
trim and still follow the affinity laws than do higher specific speed pumps. The
reasons for the deviation from the expected results are complex, related to the
fact that the impeller shrouds are not completely parallel with each other in
most pumps, that the vane exit angle is altered as the impeller is trimmed, and
other design-related reasons that are beyond the scope of this book.
Perhaps a more significant and practical use of the affinity laws is the use of
the first set of laws, Equations 2.25, 2.26, and 2.27, to determine performance
for a pump which operates at more than one speed. The increasing popularity of variable-speed pumping systems (see Chapters 6 and 7) requires the
capability of generating new H–Q and BHP curves for a pump running at a
different speed than the published or tested curve for that pump.
The following practical example serves to illustrate the value of the affinity
laws.
Example 2.6: Use of Affinity Laws to Determine
Required Speed of Pump
PROBLEM
Determine what speed the pump in Figure 2.11 should be run with full
diameter impeller to make a rating of 3000 gpm and 225 ft. The solution
to this problem cannot be determined by inspection. All that is known
from studying Figure 2.11 is that the pump must be run faster than the
speed shown on Figure 2.11, which is 1780 rpm. There is only one speed
that causes the full diameter pump curve to fall on the desired operating
point.
This type of problem may need to be solved if a pump is being transferred to a new service, or in a variable-speed pump application, where
published performance conditions exist at one speed and it is desired
to operate the pump at a higher speed to make a different rating point.
GIVEN
To solve this problem, Equations 2.25 and 2.26 are used. For this problem, N1 must be calculated, and the known terms are
Q1 = 3000 gpm
H1 = 225 ft
N2 = 1780 rpm
Hydraulics, Selection, and Curves
67
(Note that subscripts 1 and 2 above could have been reversed, with subscript 1 referring to the conditions at 1780 rpm.)
SOLUTION
This problem requires a trial and error, or iterative, solution. To solve for
N1, the steps are
1. Estimate a trial solution for N1.
2. Using the given values of Q1, H1, and N2, solve Equation 2.25 for
Q2 and Equation 2.26 for H2.
3. If the resulting data point (Q2 and H2) falls on the curve represented by the N2 speed (1780 rpm), then the trial solution is
correct. If it does not, repeat Steps 1 through 3 again with a new
trial solution.
Using the above procedure, a trial solution for N1 = 2200 rpm is proposed. Solving for Q2 and H2,
Q 2 = Q 1 × (N 2/N 1 ) (from Equation 2.25)
Q 2 = 3000 × (1780/2200) = 2427 gpm
H 2 = H 1 × (N 2 /N 1 )2 (from Equation 2..26)
H 2 = 225 × (1780/2200)2 = 147 ft
The resulting data point of 2427 gpm and 147 ft falls below the Figure
2.11 full diameter curve at 1780 rpm, so a new trial solution must be
made and the process repeated. A second trial solution of 2000 rpm for
N1 is attempted as follows:
Q 2 = Q 1 × (N 2 /N 1 ) (from Equation 2.25)
Q 2 = 3000 × (1780/2000) = 2670 gpm
H 2 = H 1 × (N 2 /N 1 )2 (from Equation 2..26)
H 2 = 225 × (1780/2000)2 = 178 ft
The resulting data point does fall on the Figure 2.11 full diameter curve at
1780 rpm, so the second trial solution for N1 is the correct one, and the pump
must be run at 2000 rpm.
If, for the second iteration, a trial solution had been chosen which was too
slow a speed (such as 1900 rpm), the resulting data point Q2 and H2 would
have been found to fall above the Figure 2.11 full diameter curve at 1780 rpm,
and a third iteration would have been required.
Once the required operating speed has been determined as shown in the
above example problem, a new set of curves at the higher speed can be developed. Figure 2.23 illustrates how pump H–Q, BHP, and efficiency curves
change with increasing speed. The H–Q and BHP curves at the higher speed
68
Pump Characteristics and Applications
260
H–Q 2000 rpm
240
100
180
90
Efficiency (%)
200
160
140
120
100
80
60
40
20
0
Brake horse power
Total head (ft)
220
H–Q 1750 rpm
80
70
Efficiency
1750 rpm
60
50
40
Efficiency
2000 rpm
30
40
20
20
10
0
0
BHP 2000 rpm
BHP 1750 rpm
0
80
160
240
320
gpm
400
480
560
FIGURE 2.23
Effects of increasing pump speed on centrifugal pump performance.
in Figure 2.23 are developed by taking arbitrary rating points (values of Q,
H, and BHP) from the lower speed curves and applying Equations 2.25, 2.26,
and 2.27 at the higher speed.
The efficiency curve for the higher speed pump is developed by taking
values of Q, H, and BHP from the higher speed pump curve, once it is developed (see preceding paragraph), and solving Equation 2.16 for efficiency η.
Note that the value of BEP efficiency stays constant, but the entire efficiency
curve shifts to the right at higher speeds, or to the left at lower speeds.
Also refer to the discussion in Chapter 3, Section III to learn how computer
software can be used to solve the above problem.
Caution should be exercised when operating pumps at higher speeds than
their published curves. Ordinarily, the manufacturer should be consulted
to verify that running the pump at the higher speed does not exceed the
design limits of the pump (i.e., the upper horsepower limit of the shaft and
the load limit of the bearings). Also, the required motor size will increase
significantly.
Another point to consider when running a pump at higher speeds than
published for that particular pump is the fact that NPSHr increases with
increasing speed, a factor not considered in the affinity law equations previously given. There is not complete agreement among manufacturers as to
the amount of variation of NPSHr with speed changes. Most pump designers
says that NPSHr varies with speed like head does, that is, with the square of
the speed change. In other words, at constant diameter:
69
Hydraulics, Selection, and Curves
NPSH 1  N 1 
=
NPSH 2  N 2 
2
(2.31)
Therefore, if a speed increase is being considered, the new NPSHr curve
should be calculated and an evaluation made of the potential of cavitation at
the higher operating speed.
Some manufacturers say that NPSHr varies with the speed ratio to the 1.5
power, rather than the 2nd power as shown by Equation 2.31. So, a conservative approach suggests that if one is planning to raise the pump speed, one
should assume that Equation 2.31 applies; if one is planning to lower the pump
speed, one should substitute a power of 1.5 in place of 2 in Equation 2.31.
IX. System Head Curves
Previous sections of this chapter have discussed the characteristic head–
capacity curve that describes the performance of a centrifugal pump with a
particular speed and impeller diameter. When pump manufacturers build
a pump, they usually select an impeller trim diameter based upon a design
operating point (total head and capacity) specified by the buyer. This rated
total head and flow, incidentally, is usually the only point on the pump curve
that the manufacturer guarantees, unless other operating points are specified by the buyer and agreed upon by the manufacturer. The rated capacity and total head usually appear on the nameplate of the pump. The only
trouble is that the pump is a rather illiterate machine and does not know
what total head and flow it is supposed to deliver. The pump only knows the
shape of its head–capacity curve based on its speed and impeller geometry,
and that it operates wherever on the H–Q curve the system tells it to operate.
This may be an operating point quite different from the rated point shown
on the nameplate. Furthermore, the point at which the pump operates on its
H–Q curve may change as system conditions change. A better understanding of these principles can be gained from a study of system head curves.
The system head curve is a plot of the total head requirement of the system
as it varies with flow. System head usually consists of a combination of static
(elevation or equivalent pressure) and frictional components, although the
relative value of these two components varies considerably from one system
to another. For example, a typical boiler feed system is composed of mostly
static head (boiler pressure is equivalent to a static head), which does not
vary with flow, whereas a typical pipe line system is composed of mostly
frictional head, which does vary with flow. As another example, a closed
loop system is composed entirely of friction head. Other examples fall everywhere between these extremes.
70
Pump Characteristics and Applications
The best way to understand system head curves is with a few illustrations.
The first example is a system that does not actually exist in the real world,
a system with no friction whatsoever. Figure 2.24 illustrates such a system,
where pipes are extremely short and extremely large in diameter, and with
no valves, so that friction losses are negligible over the entire flow range of
the pump.
The head requirement for this system, regardless of the flow, is that the
liquid must be pumped up a static distance of 19 ft, and then pumped into
the vessel that is maintained at a pressure of 100 psig ± 3 psi. Figure 2.25
shows the system head curve for this system. The system head curve, plotted on a head vs. capacity scale, is a horizontal line for this all-static system.
100 psig
± 3 psi
19 ft
FIGURE 2.24
An all-static system with no friction, used to demonstrate the development of a system head curve
257
250
243
H
(ft)
19
Variation of pump
flow rate as system
head changes
Scale change
Q (gpm)
FIGURE 2.25
System head curve for the all-static system shown in Figure 2.24, with pump H–Q curve
superimposed.
71
Hydraulics, Selection, and Curves
The lower horizontal line represents the static elevation change. The middle
of the three horizontal lines grouped together represents the system head
curve when the tank pressure is 100 psig. In that case, for all flows, the system head (with SG = 1.0) is
System head = 19 + (100 × 2.31)/1.0 = 250 ft
The other two horizontal lines grouped together in Figure 2.25 represent
the new system head curves when the pressure in the discharge vessel is,
respectively, 97 and 103 psig. When the pressure is 97 psig in the discharge
vessel, the system head is
System head = 19 + (97 × 2.31)/1.0 = 243 ft
When the discharge vessel pressure is 103 psig, the system head is
System head = 19 + (103 × 2.31)/1.0 = 257 ft
Also shown in Figure 2.25 is the H–Q curve for a pump operating in this
system. The general rule here is that the pump operates at the point of intersection of the pump H–Q curve and the system head curve. The pump seeks
the point on its H–Q curve where it delivers the head required by the system.
Figure 2.25 shows that as the pressure in the delivery vessel changes from
97 to 103 psig, the intersection of the system head curve and the pump H–Q
curve shifts, causing the pump to deliver a varying flow rate.
Figure 2.26 shows that the combination of an all-static system (no friction)
and a low-specific-speed pump (which produces a very flat H–Q curve as
discussed in Section VII above) can cause wide swings in flow with relatively
H
(ft)
Variation of pump
flow rate as system
head changes
Q (gpm)
FIGURE 2.26
The combination of a flat system head curve (no friction or very little friction) with a relatively
flat H–Q pump curve produces wide flow swings with relatively small system head variation.
72
Pump Characteristics and Applications
small changes in the system head requirement. This makes this combination
of pump and system type difficult to control and to maintain a constant flow
without the use of a flow control valve. In general, a centrifugal pump with a
steeper H–Q curve is easier to control in a system, particularly a system with
a flat or nearly flat system head curve, because variations in system head do
not produce wide swings in flow.
To continue the discussion of system head curves, consider the more realworld system shown in Figure 2.27, which includes both a static component
(elevation change of 130 ft) and a friction component (friction losses through
system piping and fittings). For this system, the system head requirement as a
function of flow is shown in Figure 2.28. The system head curve in this figure is
labeled curve X, and the pump H–Q curve is superimposed on the same scale.
The system head curve is generated by starting at H = 130 ft (the static component of head, and the total head requirement at flow rate = 0), selecting an arbitrary value of flow Q, and calculating friction losses through the entire piping
system at this flow rate. This procedure is repeated for several flow rates until a
smooth curve for the system head curve can be drawn as shown in Figure 2.28.
As before, the rule is that the intersection of the system head curve and the
pump H–Q curve, labeled point a in Figure 2.28, dictates where the pump
shown operates when running in a system described by the system head
curve labeled X in Figure 2.28.
It is possible to build a complete system head curve for a simple system
(one that does not have any branches) after only the head at one flow point
and the static head are known. The reason for this is that the friction head
130 ft
1500 ft of 6 in
SCH 40 steel pipe
FIGURE 2.27
A system composed of both static and friction head requirements.
73
Hydraulics, Selection, and Curves
X
a
H
(ft)
System curve
A
130
Q (gpm)
FIGURE 2.28
System head curve for a system composed of static and friction components, with pump H–Q
curve superimposed.
loss in a pipe line varies as the square of the flow going through the pipe
line. So once the friction loss has been calculated for a single flow rate, it
is possible to construct a table to determine the friction head loss at other
flow rates using this squared relationship. Adding that friction loss to the
static component of head gives the total head for the system at the other flow
points, and therefore other points on the system head curve are produced.
This is illustrated in Table 2.5. Given that the friction head loss is a secondorder curve, it is also possible to use a log scale on the system head curve,
which means that a straight line can be drawn from the static head at 0 flow
through the one calculated head-flow point to generate a system head curve.
This relationship of friction head varying as the square of the flow can also
be used to predict the change in head (or pressure) across a given section of
pipe, or across a component such as a heat exchanger.
In the all-static example shown in Figures 2.24 and 2.25, the system head
curve is seen to shift as the pressure in the delivery vessel changes. Similarly,
in the system shown in Figures 2.27 and 2.28, which includes friction as well
as static components, the system head curve can again change shape. This, in
TABLE 2.5
Calculation of Total System Head at Various Flow Rates
Flow Rate
(gpm)
1000
800
600
1200
Static Head
(ft)
Friction Head (ft)
Total Head (ft)
130
130
130
130
100
(800/1000)2 × 100 = 64
(600/1000)2 × 100 = 36
(1200/1000)2 × 100 = 144
230
194
166
274
Note: Calculation is simplified, once friction head is determined for a single flow
rate, since the friction head varies as the square of the flow rate.
74
Pump Characteristics and Applications
turn, causes the point of intersection with the pump H–Q curve to change. As
before, this causes the pump to operate at a different point on its H–Q curve.
Either the static component of a system head curve might shift as shown in
Figure 2.29, or the friction component of the system head curve might change,
as shown in Figure 2.30. Changes in the static component of the system head
curve might occur due to changes in the level of either the suction supply vessel or the delivery vessel, or due to variations in pressure in the supply vessel
or in the delivery vessel as illustrated in the all-static example.
Changes in the frictional component of the system head curve might be
due to several causes. These might include the buildup of solids on a filter
in the system, the throttling of a valve in the system, or the gradual buildup
of mineral deposits or corrosion products on the inside of system piping or
components. Or, the fluid might become more viscous, as in the case of oil at a
colder temperature. Any of these causes, or some combination of them, might
cause the frictional component of the system head curve to change over time,
which in turn would move the point of operation on the pump’s H–Q curve.
Actually, both the static and the frictional components of a system head
curve might change over a period of time within a system, for the reasons
discussed above, causing variations in the point of operation of the pump on
its curve. A good engineering analysis of a system during the design phase
considers the range of these swings of operation of the pump on its head–
capacity curve and considers the following for all points:
• Is the motor sized to handle the horsepower at all points?
• Is there adequate margin of NPSH at all points?
• Is the capacity of the pump at all operating points adequate to meet
process requirements?
Pump
characteristic curve
180
160
H
(ft)
System curve
low discharge
tank level
Normal system
head curve
140
120
Static elevation change
100
80
0
1
2
3
4
5
Q (gpm × 1000)
FIGURE 2.29
Change in static component of system head curve.
6
7
8
75
Hydraulics, Selection, and Curves
Pump characteristic curve
180
160
H
(ft)
140
120
Static system head
100
Normal system
head curve
80
0
1
2
3
4
5
Q (gpm × 100)
6
7
8
FIGURE 2.30
Change in friction component of system head curve.
• Does the pump operate at a point on its H–Q curve that is healthy for
the pump on a continuous basis? Consider minimum flow, bearing
loads, etc. (Refer to the discussion on minimum flow in Chapter 8,
Section III.B.)
Returning for a moment to Figure 2.21, which illustrates a drooping H–Q
curve, it should now be clear why a system head curve described by curve
A in Figure 2.21 would cause unstable operation for the pump with a drooping H–Q curve. In that case, because there are two points of intersection
of system head curve A and the pump H–Q curve, the pump would try to
operate both at point 1 and at point 2 on its H–Q curve, causing flow to surge
back and forth. On the other hand, if the pump were installed in a system
described by curve B, the pump would produce stable flow at point 3 on its
H–Q curve. This illustration of possible unstable operation resulting from
a drooping H–Q curve in conjunction with a flat system head curve is one
reason why boiler feed pumps (which operate in a system that is nearly all
static and thus are described by a nearly flat system head curve) are generally specified to have a constantly rising (not drooping) H–Q curve.
Another purpose for developing a system head curve is for analyzing
power consumption using variable-speed pumping in a system with a variable flow demand. This analysis is illustrated in Chapter 6, Section IV.
System head curves can become quite complex when they are developed
for systems with multiple branches, especially where flow is going through
more than one branch at the same time. Consider the system shown in
Figure 2.31. The system might be designed and operated in a way that has
the valves in the branches positioned such that all of the pump flow goes
76
Pump Characteristics and Applications
B
zB
3
zA
zD
2
Pump
A
C
D
zC
1
FIGURE 2.31
System with multiple branches. (From Karassik, I.J. et al., Pump Handbook, 4th Ed., McGrawHill, Inc., New York, 2008. With permission.)
through only one branch at a time (e.g., branches B and C are blocked off,
whereas branch A is open). This arrangement of a system is common, for
example, where flow in a process might go through a process stream, recycle
back to the suction supply vessel, or be pumped directly to a storage vessel.
In that case, there are actually three separate system head curves, each of
which may have a different static head requirement and a different friction
curve that varies with flow. Thus, the pump very likely operates at different
points on its H–Q curve, depending on the branch through which the flow
is directed, unless the flow is equalized by using orifices or control valves to
adjust the friction losses such that the total head loss (static plus friction) in
all three branches is equal.
If the system shown in Figure 2.31 is designed and operated such that the
flow can be directed to all three branches at the same time, a combined system head curve must be developed to determine the resulting flow produced
by the pump and to determine how the flow splits between the branches.
The system head curves for the individual branches A, B, C, and D must be
developed separately as described earlier, then combined graphically, to get
a system head curve for the entire system. Figure 2.32 illustrates how the
system curves are added geographically. Note that the system head curve
for branch D has a negative static head since the tank level is a higher elevation than junction point 1. The separate system head curves for each of the
three branches that are hydraulically in parallel (A, B, and C) are combined
by taking arbitrary values of head and adding the flow from each of the three
system curves at each head value. This produces the curve labeled A + B + C
in Figure 2.32. This part of the system is in series with branch D, and so the
77
Hydraulics, Selection, and Curves
(+)
Pump to
tal
Y
head
Junction head
vs. flow
X´
E
C
c´
a´
b
c
(A
+
zB
B
C)
A+
D
a
+
+
z B´
zA
+
A
D
C
B+
+
C
Y1´
Y1
F
Total fl ow
B
b´
zC
0
X
A
A
Pump and system heads
B´
D
h fD
zD
Rate of flow
FIGURE 2.32
Combined system head curves for multiple branch system. (From Karassik, I.J. et al., Pump
Handbook, 4th Ed., McGraw-Hill, Inc., New York, 2008. With permission.)
curve labeled A + B + C is graphically added to the system curve labeled D
by taking arbitrary values of flow and adding the head from each of the two
curves. This produces the curve labeled (A + B + C) + D, which is the system
head curve for the entire system.
The pump H–Q curve (labeled curve E in Figure 2.32) intersects with the
system head curve (A + B + C) + D at point X, and this therefore represents
the flow delivered by the pump into the system. The curve representing the
head at the junction (point 1 on Figure 2.31) as it varies with flow (labeled
curve F in Figure 2.32) is produced by observing that the head at the junction
point 1 is equal to zD (the suction head in Figure 2.31, a constant), plus the
TH of the pump (which varies with flow), less the friction head in branch D
(which also varies with flow).
Finally, the key to solving this problem is the observation that the flow
splits at the junction point 1 in such a way that the total head requirements
(static plus friction) in each of the three branches A, B, and C are equal to
each other and to the head at junction point 1. So, from point X on Figure 2.32
(the flow of the pump), a line is drawn straight down to curve F, intersecting
78
Pump Characteristics and Applications
at the point labeled Y1. This is the head at the junction when all three lines
are open and the pump is running. Drawing a line from this point Y1 straight
left until it intersects with the branch system curves labeled A, B, and C
(at points a, b, and c, respectively) produces the flow through each of these
branches. This shows that the lowest flow rate is produced in branch B, the
branch with the highest static head to overcome. If it is desired to direct more
flow through branch B, the system and probably the pump would have to be
modified to achieve this, or possibly a second pump would need to be added
at the beginning of branch B. Conversely, the highest flow rate is produced
in branch C, the branch with the lowest static head.
The above example of a multiple branch system can become very cumbersome very quickly as more branches are added to the system. The graphical
solution described above (shown in Figure 2.32) is tedious to carry out and
not practical for more than three or four branches. This illustrates one of the
benefits of computer software based piping network analysis, which is discussed more thoroughly in Chapter 3, Section III.
Another example of a common system involving multiple branches with
flow going through more than one branch simultaneously is a closed loop
chilled water system in an HVAC system. Here, balancing valves in each
parallel branch are commonly used to do a final field balance of the system
to achieve the required flow through each branch of the network. A more
accurate analysis of the branch head losses minimizes the required pressure
drop across the balance values, which saves energy.
X. Parallel Operation
Parallel operation of pumps is illustrated in Figure 2.33, where one pump,
some of the pumps, or all of the pumps can be operated at the same time. The
primary purpose of operating pumps in parallel is to allow a wider range
of flow than would be possible with a single, fixed-speed pump for systems
with widely varying flow demand. Examples of applications for parallel
pumping include municipal water supply and wastewater pumps, HVAC
system chilled water pumps, main process pumps in a variable capacity process plant, and condensate pumps in a steam power plant. Usually, there are
no more than three or four pumps operating in parallel, but large systems
may have more.
When parallel pumps are being considered for a system design, the pumps
must be carefully matched to each other and to the system to ensure that the
pumps are always operating at a healthy point on their H–Q curves, and to
ensure that the system is such that true benefits are achieved from the parallel pumping arrangement. (This does not always turn out to be the case, as
will be demonstrated shortly.)
79
Hydraulics, Selection, and Curves
Heat exchanger 1
Expansion
tank
Chilled
water
return
Heat exchanger 2
Chilled
water
supply
Pump 1
Chiller evaporator
Pump 2
Pump 3
FIGURE 2.33
Parallel operation.
To analyze a parallel pumping arrangement, it is necessary to construct
the system head curve, as explained in Section IX above. Then a combined
pump curve must be developed depicting the head–flow relationship for the
pumps while pumping in parallel. Once these two curves are constructed,
the same rule for total system flow as discussed in Section IX applies, that is,
that the total flow through the system is represented by the intersection of
the system head curve with the combined pump curve.
Figure 2.34 shows how the combined pump curve is developed, with a system having either two or three identical pumps operating in parallel. Curve
A in Figure 2.34 represents the H–Q curve for any one of the pumps. Curve B
is the combined pump curve, which represents two pumps operating at the
same time in parallel. This curve is produced by starting at the shutoff point
for curve A and moving down the H scale. At arbitrarily selected values of
head, the flow on curve A at that head is doubled, which produces a point on
curve B. This is done for about six to eight points, and then a smooth curve is
drawn through these points, creating curve B. Curve C is similarly produced
to show the combined pump curve when all three identical pumps are operated in parallel.
The system head curve is shown as curve X in Figure 2.34. The point of
intersection of pump curve A with system head curve X (labeled point a in
80
Pump Characteristics and Applications
X
H
(ft)
c
ac
b
ab
a
A
B
C
Q (gpm)
FIGURE 2.34
H–Q curve for a centrifugal pump, combined curves for two or three identical pumps operating in parallel, and system head curve.
Figure 2.34) represents the flow delivered to the system when only one pump
is operated. The point of intersection of combined pump curve B with system head curve X (labeled point b) represents the total flow delivered by the
pumps into the system when two of them are operated at the same time in
parallel. Finally, the point of intersection of curve C with system head curve
X (labeled point c) represents the total flow when all three pumps are operated in parallel.
Figure 2.34 shows that the resultant flow when two pumps are operated
in parallel in a system is not double the flow that one pump alone produces
when operating by itself, a sometimes mistaken impression. If the system
curve were completely flat (i.e., if there were no friction losses in the system),
then the flow would double when both pumps were operating. But because
the system head curve curves up due to friction in the system, two pumps
operating in parallel do not deliver double the flow of a single pump operating alone in the system. Putting the third pump on line even further diminishes the increment of flow added to the system, as Figure 2.34 illustrates.
With two identical pumps operating, the point where each pump is operating on its own H–Q curve is obtained by moving left from the intersection point of the combined two-pump curve B with the system head curve
X (point b), until the dashed line meets pump curve A. This point (labeled
point ab) represents where each pump is operating on its own H–Q curve
when these two identical pumps are operated in parallel. Similarly, the
dashed line moving left from the intersection of the combined three-pump
curve C with system head curve X (point c), intersects the pump curve A at
the point where it would operate when three pumps operate in the system
(labeled point ac). In general, with parallel pumping, each pump runs out the
furthest on its own H–Q curve when that pump operates alone in the system
(or when the fewest number of pumps allowed to be operated are running).
81
Hydraulics, Selection, and Curves
The pumps run the furthest back on their own H–Q curves when the maximum number of pumps are operated in parallel in the system.
If the system head curve is too steep, a situation that could be caused by
an undersized piping system, by some other undersized component in the
system, by foreign material trapped in a component such as check valve, or
by corrosive buildup in pipes or components that acts as a bottleneck, then it
turns out that very little benefit is achieved by operating pumps in parallel in
that system. This is illustrated in Figure 2.35, where the system head curve is
deliberately shown to be quite steep. As Figure 2.35 illustrates, the amount of
incremental flow gained by adding the second and third pumps to the system
is so small as to render the benefit from parallel pumping negligible. In this
case, capital would have been better spent, if possible, in flattening the system
head curve, by reducing bottlenecks, increasing piping sizes, or installing a
parallel pipe run to carry part of the flow, rather than in adding additional
parallel operating pumps to the system.
When considering a system with pumps in parallel, all combinations of
pumps and variations of system head curve should be considered to establish a duty cycle for the pumps, that is, for how long a period of time each
of the pumps will operate at particular points on its own H–Q curve. The
health of the pumps at these operating points should be considered with
regard to NPSH, horsepower, bearing load, and minimum flow. This will
also permit selection of the optimal pumps to minimize total energy consumed in pumping during the pumps’ life cycles.
The situation with two nonidentical pumps operating in parallel is shown
in Figure 2.36. In this figure, the curves of the two different pumps, marked
A and B, are combined in curve C, by adding the flow of the two pumps at
arbitrarily selected values of head. The combined curve is not a smooth one
X
c
b
H
(ft)
a
B
A
C
Q (gpm)
FIGURE 2.35
When the system head curve is very steep, operating a second pump in parallel with the first
produces only a marginal increase in flow.
82
Pump Characteristics and Applications
H
(ft)
X
a
b
c
A
B
C
Q (gpm)
FIGURE 2.36
Combined pump curve for two nonidentical pumps in parallel, and system head curve.
due to the way it is generated as the graphical sum of curves A and B. The
combined curve C actually follows curve A for a while, because the maximum head for curve B is lower than the maximum head for curve A. The
same rule applies as before. The intersection point of the combined curve,
and the system head curve (point c) determines the total flow of the two
pumps when they are running in parallel. The horizontal dashed line going
left from point c intersects the individual H–Q curves at the point where
each of these two pumps operates when they are operated together in parallel, labeled points a and b, respectively.
Figure 2.37 illustrates that operating two nonidentical pumps in parallel
can present problems when the two nonidentical pumps are mismatched for
the system in which they are working. Figure 2.37 is the same as Figure 2.36,
except that the system curve in Figure 2.37, labeled curve Y, is steeper than
the one shown in Figure 2.36. An attempt to run these two pumps in parallel
in the system represented by system head curve Y would mean that Pump
A would operate at the intersection point of the combined curve and the
system head curve (labeled point a in Figure 2.37), which is the same point it
would operate at when pumping alone in the system. Pump B, on the other
hand, would never be able to develop enough head to overcome the system
back pressure if both pumps were running. Thus, attempting to run the two
pumps in parallel would cause Pump B to be running at full speed but delivering no flow into the system, just as if a valve at that pump’s discharge were
completely closed. (The pump is said to be dead headed.) In addition to delivering no flow to the system, the pump would be operating at shut-off, a very
unhealthy point for continuous operation of most pumps. In fact, if the check
valve at the Pump B discharge is not sealed completely, some amount of flow
would likely be going backward through pump B.
Even if the system head curve shape in this example were such that the
intersection point between the system head curve and the combined pump
83
Hydraulics, Selection, and Curves
Y
H
(ft)
a
A
B
C
Q (gpm)
FIGURE 2.37
A mismatch occurs when attempting to operate these two nonidentical pumps in parallel in
the system shown.
curve C caused pump B to operate at a very low flow, although not at shutoff, this flow still might be below the recommended minimum continuous
flow for that pump. (Refer to the discussion in Chapter 8, Section III.B, on
minimum flow.) This type of mismatch of nonidentical pumps in a system
should obviously be avoided.
Note that a similar mismatch problem can occur with two pumps that
have the same general curve shape, but where the curve of one is higher
than the other due to differences in pump speeds or impeller diameters. If
the two pumps are significantly mismatched, this can present problems with
the lower pump operating at shut-off or below its recommended minimum
flow, even with a relatively flat system head curve.
The question is sometimes asked, while preparing the system head curve
shown in the parallel pump arrangement in Figure 2.34, how to account for
the friction losses in the branched part of the system that includes the three
pumps (see Figure 2.33). If the friction loss through each of these branches (with
approximately one third of the total flow rate passing through each branch) is
insignificant compared with the total friction loss in the rest of the system, the
splitting of the flow in these three branches can simply be ignored. On the other
hand, if the friction loss in each of the branches is a significant part of the total
system friction loss, it should not be ignored. In that case, the simplest approach
is to make an adjustment to the pump curve itself. An adjusted pump curve is
produced by calculating, at an arbitrary value of flow, the head loss through
one of the three parallel branches, and then subtracting this head from the head
on the pump curve at that flow, to produce a point on the adjusted pump curve.
This is done at several flow points, and then the adjusted pump curve can be
drawn. This adjusted pump curve is then used to build the two- and threepump combined pump curves in the same manner as before.
If the system is further complicated by the fact that the friction loss in one
of the parallel branches is significantly different from that of another branch
84
Pump Characteristics and Applications
(in the case, for example, where one of the three pumps is located a good
distance away from the other two but still piped in parallel), then this is handled exactly as described above, except the pump curves are adjusted separately. For each of the branches, an adjusted pump curve is produced, only
this time the adjusted pump curves would be different for the two pumps.
From there the combined parallel flow pump curves would be generated
using the respective adjusted pump curves, treating the pumps as if they
were two different-sized pumps.
XI. Series Operation
Series operation of pumps is illustrated in Figure 2.38. In series operation,
the discharge of one pump feeds the suction of a second pump. When two
or more pumps operate in series, the flow through all of the pumps is equal
because whatever flows through one pump must flow through the next
pump in series (provided there are no side streams).
Operating pumps in series is done for several reasons. One reason for operating pumps in series is to ensure that commercially available equipment can be
used in a particular system, while at the same time reducing system costs. For
example, if the Figure 2.38 system is a long pipe line with a large amount of friction loss across the entire pipe line, attempting to deliver the flow through the
pipe line with a single pump would result in a pump with an extremely high
head, and thus extremely high horsepower. The required horsepower might
possibly be so high, in fact, that a single pump would not be commercially
available to do the job. Some pipe lines are hundreds or even thousands of
Pump 2
Pump 1
FIGURE 2.38
Series operation.
Hydraulics, Selection, and Curves
85
miles in length, and may have a great many pumps in series as part of the pipe
line. In addition to ensuring that the pumps are of a size that are commercially
available, breaking up the line with a number of pumps in series reduces the
required design pressure for the system piping, valves, instrumentation, etc.
This same reasoning accounts for the use of several pumps in series in
a steam power plant (condensate pump, condensate booster pump, and
boiler feed pump). In theory, a single pump could pump from the condenser
through the feedwater heaters and into the boiler. However, the required
discharge pressure of this one pump would be so high that the feedwater
heaters and other system components would be outrageously expensive due
to their extremely high design pressure. Additionally, trying to do this with
a single pump would likely raise significant NPSH challenges.
Considering the above pipe line application again, another very important
reason for having multiple pumps in series is that this allows a variation of
flow through the pipe line by varying the quantity of pumps that are pumping at one time. (How to calculate the effect of serial pumping on system flow
rate is explained shortly.) Also, the use of multiple pumps in series allows the
pipe line operator flexibility to deal with a variety of pumped products having a wide range of specific gravity and viscosity.
A final application of two pumps in series is to ensure adequate available NPSH for the second pump. In this case, the first pump of two in series
might be a fairly low head pump (and thus a fairly low horsepower pump).
But, in calculating NPSHa for the second pump in series, Equation 2.19 for
NPSHa would have an additional term, namely the TH of the first pump. An
example of this is a central power station boiler feed pump, which might
operate at a speed of 6000 or 7000 rpm to develop its required head to achieve
boiler pressures. Because of this high operating speed, and assuming good
impeller design practice, this might result in an NPSHr value of as much as
100 ft for the boiler feed pump. A high NPSHa value is provided by adding a
low-speed condensate booster pump to the system.
If pumps are used in series in a system, a combined pump curve can be
generated, analogous to the combined curve discussed for parallel pumping in Section X (although the combined curve is generated differently, of
course). Again, the same general rule applies, namely that the intersection of
the combined pump curve with the system head curve determines the total
flow delivered to the system, and allows one to determine where on its own
H–Q curve each of the pumps is running.
Figure 2.39 shows how to develop a combined pump curve for two or
three identical pumps operated in series. Each pump has curve A for its H–Q
curve. Curve B, the combined curve for two pumps, is developed by starting
at zero flow, and moving to the right on the flow axis. At arbitrarily selected
flow values, the TH at each flow value is doubled, producing a point on curve
B. This process is repeated for a number of points, thereby generating curve
B. The system curve X is generated as described in Section IX, and the intersection of the system curve X with pump curve B (intersection point labeled
86
Pump Characteristics and Applications
X
c
H
(ft)
C
b
B
a
ab
ac
A
Q (gpm)
FIGURE 2.39
H–Q curve for a centrifugal pump, combined curves for two or three identical pumps operating in series, and system head curve.
b) determines the flow pumped through each pump and through the system
when two pumps are running in series. Curve C and intersection point c for
three identical pumps in series are similarly produced.
As with parallel pumping, adding pumps in series increases the amount
of flow going through the system. Pumps in series behave opposite to the
way that pumps in parallel operate, in the sense that the pumps run out the
farthest on their H–Q curves when the highest possible number of pumps
are operated in series (point ac in Figure 2.39), and they run back the farthest
on the curve when only one pump is running (point a in Figure 2.39).
Return again to Figure 2.35, which shows that when the system head curve
is very steep, operating a second pump in parallel with the first produces
only a marginal increase in flow. If these same pumps are piped in series,
rather than in parallel, they would produce a higher flow through the system. Figure 2.40 shows that two such pumps piped in series (curve D is the
combined pump curve) deliver more flow through the system (intersection point d) than three pumps operating in parallel (intersection point c).
Therefore, a general rule is that if the system head curve is relatively steep,
series pumping is probably more effective than parallel pumping for increasing the flow range of the pumps. The converse is also true, namely that for
fairly flat system head curves, parallel pumping is probably more effective
than series pumping to produce a wide flow range.
Figure 2.41 shows the combined pump curve (curve C) for two nonidentical pumps (curves A and B) operating in series. The combined curve is
generated using the same procedure as outlined above for identically sized
pumps. At the arbitrarily selected flow rates, the TH values of the two pumps
are added together to produce the combined curve data point. It is common
in pipe lines to have several differently sized pumps, allowing the operators
the widest possible range of flow and/or variation of products pumped.
87
Hydraulics, Selection, and Curves
X
H
(ft)
a
b
c
d
D
C
B
A
Q (gpm)
FIGURE 2.40
Series operation of the pumps shown in Figure 2.35 produces more flow with a steep system
head curve than parallel operation.
One final note on series pumping. If a valve in the line downstream of
a pump in a series installation is inadvertently closed completely, all of
the pumps in the line upstream of that valve move to their shutoff head,
which means that the pressure in the system downstream of the last pump
is considerably higher than normal, and possibly higher than the design
pressure of the pump casing, seal, system piping, and other components. In
X
c
H
(ft)
b
a
C
B
A
Q (gpm)
FIGURE 2.41
Combined pump curve for two nonidentical pumps in series, and system head curve.
88
Pump Characteristics and Applications
consideration of this possibility, a pressure relief system or other pressure
limits should be incorporated into the system design to keep the pressure
from exceeding design limits.
XII. Oversizing Pumps
A very common approach in sizing centrifugal pumps is to calculate capacity and total head, as described in previous sections of this chapter, and then
to adjust these terms by some sort of “fudge factor.” There are often good
reasons to adjust the design parameters of the pump in this way. Experience
has shown many engineers that the system they are designing for a process
plant today will be too small in a few years, and management will want to
get more capacity out of the plant. This is an argument for adding a fudge
factor to the capacity in sizing the pump. As for pump TH, often the engineer may not know exactly how the piping system will be built, how many
elbows will be in the system, and precisely what levels and pressures will be
present in suction and delivery vessels. Also, many piping systems are likely
to have a buildup of corrosion products on the pipe interior walls or inside
components over time, which, as discussed in Section IX, would require an
increased pump TH to maintain the same flow. This argues in favor of adding a safety factor to the calculated TH. As a pump wears, there is a small
reduction of capacity and head, so this argues in favor of adding additional
safety factor to both capacity and TH.
While these adjustments might seem reasonable, the amount of fudge factor is often arbitrarily chosen, without regard to the effect on capital and
operating costs. Also, sometimes the engineer applies a fudge factor, and
then the engineer’s supervisor will add an additional fudge factor. Then, the
pump vendor may raise the TH a bit more, just to make certain that there
is “enough pump to do the job.” Another practice sometimes followed that
might cause the fudge factor to be excessive is to take the fudge factor as a
percentage of calculated total head, rather than as a percentage of only the
friction head. Only the friction head would increase over time, for instance,
as the buildup of corrosion products causes the pipe diameter to decrease.
The static component of head would be unaffected.
The costs of adding on these fudge factors can be significant, as the following example illustrates.
Suppose a rating of 2000 gpm and 150 ft has been calculated for a pump
handling water (SG = 1.0). A fudge factor of 15% is applied to both total head
and flow to account for expected future conditions, making the rating for
which the pump is actually purchased 2300 gpm and 172.5 ft (2000 gpm and
150 ft, each multiplied by 1.15). The pump shown in Figure 2.11 is chosen
for the service. Had the original rating of 2000 gpm and 150 ft been used
Hydraulics, Selection, and Curves
89
in sizing the pump, an impeller diameter of approximately 13 3/8 in would
satisfy the condition. With this impeller diameter, the efficiency (from Figure
2.11) is 81.5%, and BHP is computed as 93 HP at the design point of 2000 gpm
and 150 ft using Equation 2.16. A 100 HP motor should provide nonoverloading service over the entire range of this pump curve.
Instead of the above selection, the 15% fudge factor is applied, and the new
rating of 2300 gpm, and 172.5 ft is used to select the pump. It turns out that
with this new rating, the same pump shown in Figure 2.11 can be used, but
with a larger impeller diameter of approximately 14 3/8 in. To fairly compare the power consumption of the pump with the two impeller diameters, a
comparison of power consumption is made at 2000 gpm. At that point on the
pump curve with a 14 3/8 in impeller, TH is about 185 ft, efficiency is about
81.9%, and BHP is computed at 114 HP. A 125 HP motor looks like it will not
provide nonoverloading service at the far right end of the curve with this
larger impeller diameter, so either the pump must operate slightly in the
service factor in this event, or else a 150 HP motor must be chosen. In either
event, the chosen motor must be larger than the 100 HP motor that satisfied
the full pump range with the smaller impeller diameter.
The decision to add the 15% adjustment to total head and flow in sizing
the pump has negative consequences for the owner of this pump for several reasons. The owner’s capital cost is higher because the pump with the
larger impeller requires a larger motor. (The pump itself probably would
cost no more because it is the same pump in either case with only a different impeller diameter.) Of greater significance, however, is the fact that from
the moment the pump is put into service, the pump with the larger impeller
consumes 114 HP to deliver 2000 gpm, vs. the 93 HP that the pump with the
smaller impeller would have consumed to deliver the same 2000 gpm. This
additional 21 HP is simply wasted energy.
It is possible to calculate the annual cost of the energy wasted if the larger
impeller diameter had been chosen by converting the 21 HP mentioned
above to kilowatts (1 HP = 0.746 KW, 21 HP = 15.7 KW), then multiplying
the result by the owner’s cost of power in dollars per kilowatt-hour, and by
the anticipated number of hours of operation of the pump in a year. Cost of
power varies widely from location to location, and depends on the amount
of consumption and several other factors. In the United States, most industrial and commercial users pay from $0.05 to $0.14/kW-h. Assuming a cost of
power of $0.10/kW-h, and 6000 hours of operation per year (roughly operating 70% of full-time operation), the annual cost to the pump owner of the
energy wasted by choosing the larger impeller diameter is
(15.7 KW) × ($0.10/kW-h) × (6000 HR/year) = $9420/year
The annual cost to the owner for the energy wasted by choosing the larger
impeller can be nearly as much as the capital cost of the pump itself! Clearly,
90
Pump Characteristics and Applications
it does not make good economic sense to waste such a large amount of energy
beginning from the day the pump is commissioned, merely in anticipation of
the need for more capacity and total head some years in the future.
A third negative consequence of excessively oversizing the pump is that
this oversized pump requires more throttling to achieve lower flow rates
than would a smaller pump or a pump with a smaller impeller. The lowest
flow rate required for this system (at start-up or recycle, for example) would
then be a smaller percentage of BEP flow than would be the case if the chosen
pump had a smaller impeller diameter. Refer to related discussions on this
topic in the Chapter 8, Section III.B, discussion of minimum flow.
Oversizing a pump can lead to significant negative consequences, as the
preceding discussion has summarized. On the other hand, to totally ignore
the possibility of needing higher flow and total head from a pump at some
time in the future is not wise either. An alternative that makes sense in the
example shown here would be to buy the pump with the smaller 13 3/8-indiameter impeller, but spend the extra money to buy the larger 125 HP motor
for the pump. Then, 5 years or so later when the higher total head and flow
are required, all that is necessary is to install the larger impeller. The bigger
motor is purchased in the beginning so that it will not have to be changed
out at a later date to accommodate the anticipated larger impeller. Sometimes,
retrofitting a larger motor can be troublesome if it requires a larger pump
bedplate or different electrical requirements. It may be easier to simply put in
the larger motor right away. The difference in capital cost for the larger motor
size may not be that significant. Also, the 125 HP motor in this example operating at a lower percentage of full load generally should not suffer a loss of
efficiency compared with the 100 HP motor. If the larger motor is not installed
at the outset, the electrical infrastructure (cable size, starter, pump bedplate,
etc.) should be sized at the outset to handle the larger motor so that switching
to a larger motor later in the life of the pump (at the time a larger impeller is
installed) does not require significant electrical system modifications.
One last lesson this example teaches is the wisdom of not choosing the
maximum impeller diameter for an application, allowing the possibility of
putting in a larger impeller at a later date to stretch the capacity and total
head of the pump if needed. Using good pump selection practices, the impeller chosen for a particular application should normally not exceed 90% to
95% of the maximum size.
XIII. Pump Speed Selection
The operating speed for each of the centrifugal pumps considered in
this chapter has been treated as a given. Practically speaking, for a given
application, there may be a range of speeds considered by the user and
Hydraulics, Selection, and Curves
91
offered by the manufacturer for pumps that are suitable for the service.
This leads to the question of how the pump manufacturer determines
what are the appropriate speeds to offer for a particular pump type and
size. Or alternatively, given a particular capacity and head requirement,
how does one consider which speeds should be considered for a given
application?
There are several criteria that should be considered in combination, as
follows.
A. Suction Specific Speed
Suction specific speed was introduced in Section VII. Using the recommended value of 8500 for suction specific speed as a criterion, Equation 2.24
can be used to choose a maximum operating speed for a given application.
This is done by taking the NPSHa from the planned installation and subtracting a reasonable margin between NPSHa and NPSHr (Section VI.E) to
arrive at a value of NPSHr to use in Equation 2.24. Using the design capacity
Q, Equation 2.24 can then be solved for N, the maximum speed. A motor
with nominal speed below this calculated maximum would then be chosen.
This will not be the only choice for pump speed, but will generally result in
the smallest sized pump that will suit the application, and the lowest first
cost alternative. Other criteria, as described in the following sections, might
dictate the use of a slower pump speed.
B. Shape of Pump Performance Curves
The speed chosen for a particular application (along with the number of
stages if the pump is multistage) affects the specific speed Ns of the pump
chosen. As Section VII demonstrated, this has an effect on the shape of the
pump H–Q and BHP curves. It may be desirable, for example, for a particular
application to avoid having a high head or high horsepower at shutoff, both
of which could result if a pump having the higher value of N, and thus higher
Ns, is chosen. Reasons for not wanting the higher specific speed pump might
be due to the greater control problems that the steeper pump curve could
cause, or concerns with start-up motor overload problems with two pumps
in parallel. These considerations can result in the choice of a speed that is less
than the maximum one indicated by the criterion of suction specific speed
discussed in Section III.A above.
C. Maximum Attainable Efficiency
Refer to Figure 2.42, which shows the range of efficiencies to be expected for
centrifugal pumps as a function of specific speed and capacity. Although
these are only theoretical values of the maximum efficiency obtainable, they
do represent what is generally obtainable in the centrifugal pump industry
92
Pump Characteristics and Applications
1.0
0
10
20
NSm
30
40
η∞
50
10,000 gpm (2300 m3/h)
3000 gpm (680)
1000 gpm (230)
500 gpm (115)
300 gpm (70)
200 gpm (45)
100 gpm (23)
0.9
0.8
0.7
50 gpm (11)
30 gpm (7)
0.6
η 0.5
0.4
10 gpm (2.3 m3/h)
5 gpm (1.1 m3/h)
0.3
0.2
0.1
0
0
500
1000
1500
2000
2500
3000
FIGURE 2.42
Pump efficiency as a function of specific speed and capacity. (From Karassik, I.J. et al., Pump
Handbook, 4th Ed., McGraw-Hill, Inc., New York, 2008. With permission.)
with good design and manufacturing practices. The curves shown in Figure
2.42 each represent a different best efficiency point capacity. This figure
shows that, in general, the higher the BEP capacity of the pump, the greater
the expected best efficiency of that pump. And, for a given pump capacity,
the larger the design specific speed of the pump, the greater the expected
pump efficiency (up to a point at least). This figure helps explain why the
efficiency of very small centrifugal pumps is relatively low, whereas that of
very large centrifugal pumps may be greater than 90%.
Figure 2.42 suggests that there is an optimum specific speed range to
achieve the highest BEP for the pump chosen for a particular application.
Chapter 6, Section II, illustrates in several examples how the selection of a
higher operating speed (or number of stages) affects specific speed, which in
turn can affect the efficiency of the pump chosen for the application.
Another, more detailed approach at predicting pump efficiency is taken
by the Hydraulic Institute, as shown In Figures 2.43, 2.44, and 2.45. Figure
2.43 shows the generally attainable efficiency levels of centrifugal pumps at
BEP with maximum diameter impellers when pumping clear water at 85°F,
as a function of the flow rate at the BEP. Note that this figure shows different
curves of predicted efficiency for various pump types, with the pump types
being summarized in Table 2.6.
Figure 2.44 shows a reduction to be made from the predicted efficiency in
Figure 2.43, as a function of the specific speed of the pump. Note that there
93
Hydraulics, Selection, and Curves
95
90
G
85
80
J
Efficiency at optimum specific speed
75
V
70
65
60
D
H
Note:
1. Charts depict the generally attainable
efficiency levels of centrifugal pumps at BEP
with maximum diameter impellers when
pumping clear water at 85 °F.
2. Charts relate to industrial class pumps,
manufactured and tested in accordance with
recognized industry standards.
F
E
55
50
45
40
C
B
A
35
30
100
1000
Rate of flow (gpm)
10,000
100,000
FIGURE 2.43
Optimum normally attainable efficiency chart (U.S. customary units). (Courtesy of the
Hydraulic Institute, Parsippany, NJ; www.pumps.org.)
are two efficiency correction curves in Figure 2.44, one curve for vertical
turbine, mixed flow, and propeller pumps, and a second curve for all other
pump types.
Finally, Figure 2.45 shows the deviation from the normally attainable efficiency that would be predicted by Figures 2.43 and 2.44, with this deviation based on the Hydraulic Institute’s survey of offerings from U.S. pump
manufacturers.
94
Pump Characteristics and Applications
5.0
4.5
Note:
1. Values for Ns for double suction pumps are based
on one half of flow rate.
2. Values for Ns for multistage pumps are based on
head per stage.
Efficiency correction (points)
4.0
3.5
All but
“V”
3.0
2.5
“V”
2.0
1.5
1.0
0.5
0.0
100
1000
10,000
100,000
Specific speed (in US units)
Deviation as a percent of generally attainable efficiency
FIGURE 2.44
Efficiency reduction due to specific speed. (Courtesy of the Hydraulic Institute, Parsippany, NJ;
www.pumps.org.)
14
12
10
8
6
4
Positive
2
Generally attainable efficiency
0
–2
–4
Maximum
Negative
–6
Minimum
–8
–10
–12
–14
10
100
1000
Rate of flow (US gpm)
10,000
100,000
FIGURE 2.45
Deviation from normally attainable efficiency. (Courtesy of the Hydraulic Institute, Parsippany,
NJ; www.pumps.org.)
95
Hydraulics, Selection, and Curves
TABLE 2.6
Pump Types Included in Figure 2.43
Category
A
B
C
D
E
F
G
H
J
V
Pump Type Description
Slurry pump, end suction
Solids—handling, end suction
Submersible sewage, end suction
Paper stock pump, end suction
Multistage, axially and radially split
ASME B73.1, API 610, small end suction
End suction—large (>5000 gpm)
API double suction
Double suction
Vertical turbine, mixed flow, and propeller (bowl efficiency)
Note that at the time of the writing of this edition, standards are being
formulated in both the United States and Europe to regulate minimum efficiencies for certain centrifugal pumps.
D. Speeds Offered by Manufacturers
In general, the speed choices to consider are limited to those that are commercially offered by the manufacturers of that pump type. In addition to the
criteria described above, other limitations on maximum speed offered by the
manufacturer for a given pump type might include maximum available shaft
size for the pump, largest bearing system offered for a given pump, casing
pressure limits, or other mechanical design limitations. Speeds higher than
those published by the manufacturer should only be considered after receiving confirmation from the manufacturer of the acceptability of the speed
based on these mechanical design considerations. Consideration of operating speeds other than the synchronous motor speeds offered by the manufacturer (e.g., through the use of variable speed drives, which are discussed
in Chapter 6, Section IV) should include a verification that the planned operating speed is sufficiently far away (at least 20%) from the pump natural frequency and the shaft critical speed. Operating a pump at a speed too close
to the natural frequency of the pump or the rotating element produces unacceptable vibration levels, a condition known as resonance. This is more likely
to be a potential problem with larger equipment, such as vertical turbine
pumps (see Chapter 4, Section XI).
E. Prior Experience
The selecting engineer should consider the prior operation and maintenance
experience with similar pumps in the speeds being considered at the facility where the pump will run. Prior problems, or lack thereof, with pumps in
96
Pump Characteristics and Applications
similar or related applications may signal the acceptability of a higher speed
choice for a particular application. In looking for examples of prior experience, consideration should include the liquid being pumped (especially the
amount of abrasives or level of corrosiveness), past success with keeping similarly sized pumps at the higher speed properly aligned and balanced, and
the facility’s experience with seals and bearings with similarly sized pumps
at the higher operating speed. Also, higher speed motors generally produce
higher noise levels, which might be a consideration if the pumps are operating in a location where people are present.
3
Special Hydraulic Considerations
I. Overview
This chapter discusses some special hydraulic considerations that are
beyond standard application problems. It includes a discussion of viscosity,
describing terms and units used to measure viscosity. An example shows the
reader how to adjust centrifugal pump performance when handling viscous
liquids, and how to judge whether or not it makes sense to even try to use a
centrifugal pump with a highly viscous liquid.
Section III of this chapter is devoted to a discussion of using computer
software to design, analyze, and optimize piping systems and to select
pumps.
Section IV describes the importance of proper piping design and installation to avoid cavitation, air entrainment, and other field problems.
Field testing—that is, producing a pump’s performance curve by measuring total head, flow, and power for an installed pump—is demonstrated in Section VI. The usefulness of field testing in pump preventive
maintenance and troubleshooting is explained. Some practical considerations, such as what is the best way to measure flow and total head in
the field and what types of devices are available to do this, are included.
This includes a brief description of the most common types of field testing
equipment, as well as a summary of the merits and shortcomings for each
type of device.
II. Viscosity
The viscosity of a liquid can be thought of as the resistance of one layer of
liquid to movement against another layer, or the internal friction within a
liquid that resists a shear force. Viscous liquids can be found in everyday
life, with examples including food items like syrup and catsup, personal care
products such as shampoo and lotion, and automobile engine oil.
© 2008 Taylor & Francis Group, LLC
1
2
Pump Characteristics and Applications
There are several categories of liquid based on their behavior when subjected to shear forces. Newtonian liquids are “well behaved,” that is, they act
like water. Their viscosity remains constant as rate of shear increases.
Some liquids, called pseudo-plastic liquids, have a decreasing viscosity with
increasing rates of shear. These liquids, examples being grease and paint, are
not difficult to pump because they tend to thin out when subjected to the
rather high rates of shear inside a centrifugal pump.
Plastic liquids act just like pseudo-plastic ones, except that a certain force
must be applied before the liquid begins to move. An example of a plastic
liquid is tomato catsup.
Dilatant liquids behave in just the opposite way of pseudo-plastic liquids.
That is, the viscosity increases in these liquids with increasing rates of shear.
These liquids, examples being clay slurries and candy compounds, cannot
generally be pumped with centrifugal pumps.
Viscosity is usually expressed in units of centipoise, centistokes, or SSU
(Seconds Saybolt Universal). These terms are related by the following formulae:
Centistokes = centipoise/SG
(3.1)
Centistokes = 0.22 SSU – 180/SSU
(3.2)
SSU = centistokes × 4.635 (for >70 centistokes)
(3.3)
There are many other units used to express viscosity of liquids. These have
been developed by different industries to describe their particular liquid
products. Table 3.1 can be used for converting from other viscosity units into
centistokes or SSU.
Table 3.2 lists the viscosity of some common liquids, in units of SSU and
centistokes, as well as the specific gravity of these liquids. Note that the values of viscosity are given at several temperatures, with viscosity decreasing
as temperature increases. Note also that the range of viscosities in Table 3.2
is quite wide, ranging from less than 100 to 500,000 SSU.
If a centrifugal pump is being considered for pumping a viscous liquid,
the pump performance curve must be adjusted for the effect of the viscosity.
In general, pumping a viscous liquid causes some reduction in both head
and capacity and usually a significant reduction in efficiency. These general
trends are illustrated in Figure 3.1.
Figure 3.2 shows a chart that can be used for correcting pump performance
when handling a viscous liquid. This chart is restricted for use only with
Newtonian liquids and radial flow pumps. Note that this chart is considered
by some manufacturers to be overly conservative in estimating the corrected
performance of a pump in a viscous application.
To make a preliminary pump selection for a viscous application, enter the
bottom of the chart in Figure 3.2 with the required viscous pump capacity (in
gpm). Then move vertically on the chart to the sloped line representing the
© 2008 Taylor & Francis Group, LLC
1.00
2.56
4.30
7.40
10.3
13.1
15.7
18.2
20.6
32.1
43.2
54.0
65.0
87.6
110.0
132
154
176
35
40
50
60
70
80
90
100
150
200
250
300
400
500
600
700
800
Kinematic
Viscosity
(Centistokes)
31
Seconds
Saybolt
Universal
(SSU)
Viscosity Conversion
TABLE 3.1
© 2008 Taylor & Francis Group, LLC
81.0
71.1
61.4
51.6
41.9
32.5
28.0
23.5
19.30
15.24
14.44
13.70
12.95
—
—
—
—
—
Seconds
Saybolt
Furol
(SSF)
677
592
508
423
338
254
212
170
128
85.6
77.6
69.2
60.9
52.3
44.3
36.2
32.1
29
Seconds
Redwood 1
(Standard)
73.8
64.6
55.4
46.2
37.1
28.0
23.45
18.90
14.48
10.12
9.30
8.44
7.60
6.77
5.83
5.10
—
—
Seconds
Redwood 2
(Admiralty)
23.35
20.45
17.50
14.60
11.70
8.79
7.35
5.92
4.48
3.02
2.73
2.45
2.17
1.88
1.58
1.31
1.16
1.00
Degrees
Engler
35.2
40.3
47.0
56.4
70.8
95
114
144
195
307
348
404
483
618
838
1440
2420
6200
Degrees
Barbey
120
106
92
79
66
52.5
46
40
—
—
—
—
—
—
—
—
—
—
Seconds
Parlin
Cup #7
39
35
30
25
21
15
—
—
—
—
—
—
—
—
—
—
—
—
Seconds
Parlin
Cup #10
9.8
9.0
8.5
7.8
7.2
6.0
—
—
—
—
—
—
—
—
—
—
—
—
Seconds
Parlin
Cup #15
4.1
3.9
3.6
3.4
3.2
3.0
—
—
—
—
—
—
—
—
—
—
—
—
Seconds
Parlin
Cup #20
74
67
58
50
42
30
—
—
—
—
—
—
—
—
—
—
—
—
Seconds
Ford
Cup #3
(continued)
50
45
40
34
28
20
—
—
—
—
—
—
—
—
—
—
—
—
Seconds
Ford
Cup #4
Special Hydraulic Considerations
3
© 2008 Taylor & Francis Group, LLC
220
330
440
550
660
880
1100
1320
1540
1760
1980
2200
3300
4400
1500
2000
2500
3000
4000
5000
6000
7000
8000
9000
10,000
15,000
20,000
2000
1500
1000
900
800
700
600
500
400
300
250
200
150
100.7
91.0
Seconds
Saybolt
Furol
(SSF)
18,400
13,700
8460
7620
6770
5920
5080
4230
3380
2540
2120
1690
1270
896
762
Seconds
Redwood 1
(Standard)
—
—
921
829
737
645
553
461
368
276
230
184.2
138.2
92.1
83.0
Seconds
Redwood 2
(Admiralty)
584
438
292
263
233.5
204.5
175
146
117.0
87.60
73.0
58.40
43.80
29.20
26.30
Degrees
Engler
1.40
2.50
2.82
3.13
3.52
4.03
4.70
5.64
7.05
9.4
11.3
14.1
18.7
28.2
31.3
Degrees
Barbey
—
—
—
—
—
—
—
—
—
—
—
—
—
149
135
Seconds
Parlin
Cup #7
860
650
430
387
344
300
258
215
172
129
108
86
65
43
41
Seconds
Parlin
Cup #10
203
147
96
86
76
67
57
47
37
28.5
24
19.5
15.2
11.5
10.7
Seconds
Parlin
Cup #15
Source: Hydraulic Institute, Engineering Data Book, 2nd edition. Parsippany, NJ, 1990; www.pumps.org. With permission.
198
1000
Kinematic
Viscosity
(Centistokes)
900
Seconds
Saybolt
Universal
(SSU)
Viscosity Conversion
TABLE 3.1 (Continued)
70
53
35
32
29
25
22
18
14
11
9
7.5
6.3
4.5
4.3
Seconds
Parlin
Cup #20
1715
1280
850
780
680
600
520
425
337
258
218
172
132
90
82
Seconds
Ford
Cup #3
1150
860
575
520
465
410
350
290
230
172
147
118
90
62
57
Seconds
Ford
Cup #4
4
Pump Characteristics and Applications
5
Special Hydraulic Considerations
TABLE 3.2
Viscosity of Common Liquids
Liquid
aSpecific
Gravity at 60°F
Viscosity
SSU
Centistokes
At °F
0.27–0.32
70
68.6
100
68
70
100
65
100
100
68
Glycerine (100%)
1.37 to 1.49 at
70°F
1.26 at 68°F
Hydrochloric acid (31.5%)
Mercury
1.05 at 68°F
13.6
Phenol (carbolic acid)
Silicate of soda
Sulfuric acid (100%)
0.95 to 1.08
40 Baume
42 Baume
1.83
65
365
637.6
75.7
648
176
1.9
0.118
0.11
11.7
79
138
14.6
Glycol:
Propylene
Triethylene
Diethylene
Ethylene
1.038 at 68°F
1.125 at 68°F
1.12
1.125
240.6
185.7
149.7
88.4
52
40
32
17.8
70
70
70
70
Fish and animal oils:
Bone oil
0.918
Cod oil
0.928
Lard
0.96
Lard oil
0.912 to 0.925
220
65
150
95
287
160
190 to 220
112 to 128
140
90
230
130
110
78
163 to 184
97 to 112
47.5
11.6
32.1
19.4
62.1
34.3
41 to 47.5
23.4 to 27.1
29.8
18.2
49.7
27.5
23.0
15.2
35 to 39.6
19.9 to 23.4
130
212
100
130
100
130
100
130
100
130
100
130
100
130
100
130
165 to 240
90 to 120
240 to 400
120 to 185
35.4 to 51.9
100
18.2 to 25.3
130
51.9 to 86.6
100
25.3 to 39.9
130
(continued)
Freon
Menhaden oil
0.933
Neatsfoot oil
0.917
Sperm oil
0.883
Whale oil
0.925
Mineral oils:
Automobile crankcase oils
SAE 10
0.880 to 0.935b
SAE 20
0.880 to 0.935a
© 2008 Taylor & Francis Group, LLC
2950
813
6
Pump Characteristics and Applications
TABLE 3.2 (Continued)
Viscosity of Common Liquids
Liquid
aSpecific
Gravity at 60°F
SAE 30
0.880 to 0.935
SAE 40
0.880 to 0.935b
SAE 50
0.880 to 0.935b
SAE 60
0.880 to 0.935b
SAE 70
0.880 to 0.935b
SAE 10W
SAE 20W
0.880 to 0.935b
0.880 to 0.935b
b
Automobile transmission lubricants:
SAE 80
0.880 to 0.935b
SAE 90
0.880 to 0.935b
SAE 140
0.880 to 0.935b
SAE 250
0.880 to 0.935b
Crude oils:
Texas, Oklahoma
0.81 to 0.916
Wyoming, Montana
0.86 to 0.88
California
0.78 to 0.92
Pennsylvania
0.8 to 0.85
Viscosity
SSU
Centistokes
At °F
400 to 580
185 to 255
580 to 950
255 to 80
950 to 1600
80 to 105
1600 to 2300
105 to 125
2300 to 3100
125 to 150
5000 to 10,000
10,000 to 40,000
86.6 to 125.5
39.9 to 55.1
125.5 to 205.6
55.1 to 15.6
205.6 to 352
15.6 to 21.6
352 to 507
21.6 to 26.2
507 to 682
26.2 to 31.8
1100 to 2200
2200 to 8800
100
130
100
130
100
210
100
210
100
210
0
0
100,000 max
800 to 1500
300 to 500
950 to 2300
120 to 200
Over 2300
Over 200
22,000 max
173.2 to 324.7
64.5 to 108.2
205.6 to 507
25.1 to 42.9
Over 507
Over 42.9
0
100
130
130
210
130
210
40 to 783
34.2 to 210
74 to 1215
46 to 320
40 to 4840
34 to 700
46 to 216
38 to 86
4.28 to 169.5
2.45 to 45.3
14.1 to 263
6.16 to 69.3
4.28 to 1063
2.4 to 151.5
6.16 to 46.7
3.64 to 17.2
60
100
60
100
60
100
60
100
Diesel engine lubricating oils (based on average midcontinent paraffin base):
Federal Specification
0.880 to 0.935b
165 to 240
35.4 to 51.9
100
No. 9110
90 to 120
18.2 to 25.3
130
Federal Specification
0.880 to 0.935b
300 to 410
64.5 to 88.8
100
No. 9170
140 to 180
29.8 to 38.8
130
Federal Specification
0.880 to 0.935b
470 to 590
101.8 to 127.8
100
No. 9250
200 to 255
43.2 to 55.1
130
Federal Specification
0.880 to 0.935b
800 to 1100
173.2 to 238.1
100
No. 9370
320 to 430
69.3 to 93.1
130
Federal Specification
0.880 to 0.935b
490 to 600
106.1 to 129.9
130
No. 9500
92 to 105
18.54 to 21.6
210
(continued)
© 2008 Taylor & Francis Group, LLC
7
Special Hydraulic Considerations
TABLE 3.2 (Continued)
Viscosity of Common Liquids
Liquid
aSpecific
Gravity at 60°F
Viscosity
SSU
Centistokes
At °F
Diesel fuel oils:
No. 2 D
0.82 to 0.95b
32.6 to 45.5
39
2 to 6
1 to 3.97
100
130
No. 3 D
0.82 to 0.95b
No. 4 D
0.82 to 0.95b
No. 5 D
0.82 to 0.95b
45.5 to 65
39 to 48
140 max
70 max
400 max
165 max
6 to 11.75
3.97 to 6.78
29.8 max
13.1 max
86.6 max
35.2 max
100
130
100
130
122
160
Fuel oils:
No. 1
0.82 to 0.95b
No. 2
0.82 to 0.95b
No. 3
0.82 to 0.95b
No. 5A
0.82 to 0.95b
No. 5B
0.82 to 0.95b
No. 6
0.82 to 0.95b
Fuel oil—Navy
Specification
Fuel oil—Navy II
0.989 maxb
34 to 40
32 to 35
36 to 50
33 to 40
35 to 45
32.8 to 39
50 to 125
42 to 72
125 to 400
72 to 310
450 to 3000
175 to 780
110 to 225
63 to 115
1500 max
480 max
Gasoline
0.68 to 0.74
Gasoline (natural)
Gas oil
76.5° API
28° API
73
50
2.39 to 4.28
2.69
3.0 to 7.4
2.11 to 4.28
2.69 to 0.584
2.06 to 3.97
7.4 to 26.4
4.91 to 13.73
26.4 to 86.6
13.63 to 67.1
97.4 to 660
37.5 to 172
23 to 48.6
11.08 to 23.9
324.7 max
104 max
0.46 to 0.88
0.40 to 0.71
0.41
13.9
7.4
70
100
70
100
100
130
100
130
100
130
122
160
122
160
122
160
60
100
68
70
100
115 max
65 max
35
32.6
24.1 max
11.75 max
2.69
2
70
100
68
100
Insulating oil:
Transformer, switches, and
circuit breakers
Kerosene
1.0 max
0.78 to 0.82
Machine lubricating oil (average Pennsylvania paraffin base):
Federal Specification No. 8
0.880 to 0.935b
112 to 160
70 to 90
© 2008 Taylor & Francis Group, LLC
23.4 to 34.3
100
13.1 to 18.2
130
(continued)
8
Pump Characteristics and Applications
TABLE 3.2 (Continued)
Viscosity of Common Liquids
aSpecific
Gravity at 60°F
Viscosity
SSU
Centistokes
At °F
160 to 235
90 to 120
235 to 385
120 to 185
385 to 550
185 to 255
34.3 to 50.8
18.2 to 25.3
50.8 to 83.4
25.3 to 39.9
83.4 to 119
39.9 to 55.1
100
130
100
130
100
130
140 to 190
86 to 110
190 to 220
110 to 125
100
77
29.8 to 41
17.22 to 23
41 to 47.5
23 to 26.4
20.6
14.8
100
130
100
130
130
160
0.91 Average
400 to 440
185 to 205
86.6 to 95.2
39.9 to 44.3
100
130
Vegetable oils:
Castor oil
0.96 at 68 F
China wood oil
0.943
Coconut oil
0.925
Corn oil
0.924
1200 to 1500
450 to 600
1425
580
140 to 148
76 to 80
135
54
176
100
143
93
200
115
221
125
195
112
250
145
1500
600
500 to 20,000
1000 to 50,000
Liquid
Federal Specification
No. 10
Federal Specification
No. 20
Federal Specification
No. 30
Mineral lard cutting oil:
Federal Specification
Grade 1
Federal Specification
Grade 2
Petrolatum
Turbine lubricating oil:
Federal Specification
(Penn Base)
0.880 to 0.935
b
0.880 to 0.935b
0.880 to 0.935b
**0.825
Cotton seed oil
0.88 to 0.925
Linseed oil, raw
0.925 to 0.939
Olive oil
0.912 to 0.918
Palm oil
0.924
Peanut oil
0.920
Rape seed oil
0.919
Rosin oil
0.980
Rosin (wood)
1.09 average
© 2008 Taylor & Francis Group, LLC
259.8 to 324.7
100
97.4 to 129.9
130
308.5
69
125.5
100
29.8 to 31.6
100
14.69 to 15.7
130
28.7
130
8.59
212
37.9
100
20.6
130
30.5
100
18.94
130
43.2
100
24.1
130
47.8
100
26.4
130
42
100
23.4
130
54.1
100
31
130
324.7
100
129.9
130
108.2 to 4400
200
216.4 to 11,000
190
(continued)
9
Special Hydraulic Considerations
TABLE 3.2 (Continued)
Viscosity of Common Liquids
Liquid
aSpecific
Gravity at 60°F
Viscosity
SSU
Centistokes
At °F
184
110
165
96
33
32.6
39.6
23
35.4
19.64
2.11
2.0
100
130
100
130
60
100
Sesame oil
0.923
Soya bean oil
0.927 to 0.98
Turpentine
0.86 to 0.87
Sugar, syrups, molasses, etc.
Corn syrups
1.4 to 1.47
5000 to 500,000
1500 to 60,000
1100 to 110,000
324.7 to 13,200
100
130
Glucose
1.35 to 1.44
35,000 to 100,000
4000 to 11,000
340
1300 to 23,000
700 to 8000
6400 to 60,000
3000 to 15,000
17,000 to 250,000
6000 to 75,000
7700 to 22,000
880 to 2420
73.6
281.1 to 5070
151.5 to 1760
1410 to 13,200
660 to 3300
2630 to 5500
1320 to 16,500
100
150
100
100
130
100
130
100
130
230
92
310
111
440
148
650
195
1000
275
1650
400
2700
640
5500
1,100
10,000
2000
49.7
18.7
67.1
23.2
95.2
31.6
140.7
42.0
216.4
59.5
364
86.6
595
138.6
1210
238
2200
440
70
100
70
100
70
100
70
100
70
100
70
100
70
100
70
100
70
100
Honey (raw)
Molasses “A”
(first)
Molasses “B”
(second)
Molasses “C”
(blackstrap or final)
1.40 to 1.46
1.43 to 1.48
1.46 to 1.49
Sucrose solutions (sugar syrups):
60 Brix
1.29
62 Brix
1.30
64 Brix
1.31
66 Brix
1.326
68 Brix
1.338
70 Brix
1.35
72 Brix
1.36
74 Brix
1.376
76 Brix
1.39
Tars:
Tar-coke oven
1.12+
© 2008 Taylor & Francis Group, LLC
3000 to 8000
650 to 1400
600 to 1760
71
140.7 to 308
100
(continued)
10
Pump Characteristics and Applications
TABLE 3.2 (Continued)
Viscosity of Common Liquids
Liquid
aSpecific
Gravity at 60°F
Viscosity
SSU
Centistokes
At °F
Tar-gas house
1.16 to 1.30
15,000 to 300,000
2000 to 20,000
3300 to 66,000
440 to 4400
70
100
Road tar:
Grade RT-2
1.07+
Grade RT-4
1.08+
Grade RT-6
1.09+
Grade RT-8
1.13+
Grade RT-10
1.14+
Grade RT-12
1.15+
Pine tar
1.06
200 to 300
55 to 60
400 to 700
65 to 75
1000 to 2000
85 to 125
3000 to 8000
150 to 225
20,000 to 60,000
250 to 400
114,000 to 456,000
500 to 800
2500
500
43.2 to 64.9
8.77 to 10.22
86.6 to 154
11.63 to 14.28
216.4 to 440
16.83 to 26.2
660 to 1760
31.8 to 48.3
4400 to 13,200
53.7 to 86.6
25,000 to 75,000
108.2 to 173.2
559
108.2
122
212
122
212
122
212
122
212
122
212
122
212
100
132
Corn starch solutions:
22 Baume
1.18
24 Baume
1.20
25 Baume
1.21
150
130
600
440
1400
800
32.1
27.5
129.8
95.2
303
173.2
70
100
70
100
70
100
2500 to 10,000
1100 to 3000
56
550 to 2200
238.1 to 660
9.07
1.13
313
143
1.13
0.55
100
130
212
68
68
100
60
130
Miscellaneous
Ink-printers
1.00 to 1.38
Tallow
Milk
Varnish-spar
0.918 average
1.02 to 1.05
0.9
Water-fresh
1.0
1425
650
Source: Hydraulic Institute, Engineering Data Book, 2nd edition. Parsippany, NJ, 1990; www.
pumps.org. With permission.
a Unless otherwise noted.
b Depends on origin or percent and type of solvent.
© 2008 Taylor & Francis Group, LLC
11
Efficiency
NPSHr
Power
Total head
Special Hydraulic Considerations
Flow rate
Flow rate
Water
Viscous liquid
FIGURE 3.1
General effects of viscosity on centrifugal pump performance.
required head per stage of the pump, expressed in feet of the viscous liquid.
For example, for a two-stage pump with a viscous rating of 500 gpm and 400 ft,
enter the chart at 500 gpm, and move up to the 200-ft line.
Next, move horizontally on the chart (either to the right or to the left) to the
sloped line representing the viscosity of the liquid (shown in Figure 3.2 in
both centistokes and in SSU). Finally, move vertically to the top of the chart,
intersecting the correction factor curves Cη, CQ, and CH, which represent the
correction factors for efficiency, capacity, and head, respectively. There are
four separate head correction factors in Figure 3.2; but for the preliminary
pump selection, use the correction factor curve labeled 1.0 × QNW.
The desired capacity and head of the pump when handling the viscous liquid are then divided by the respective correction factors CQ and CH obtained
above (in decimal form), to determine the equivalent water rating of the
pump, allowing a pump size to be selected. Once the pump size and impeller diameter are selected, the BEP of this pump with the chosen impeller
diameter is used to re-enter the Figure 3.2 correction chart to obtain revised
correction factors. These final correction factors are then used to create a new
pump curve for the pump handling the viscous liquid.
The following example of selecting a pump for a viscous application illustrates the use of the Figure 3.2 performance correction chart.
Example 3.1: Selecting a Pump for a Viscous Service,
Calculating Effect of Viscosity on Performance,
and Developing New Pump Curves
PROBLEM
Select a pump to deliver 750 gpm at 100 ft total head of a liquid having a
viscosity of 1000 SSU and a specific gravity of 0.90 at the pumping temperature. (From Table 3.2, SAE 40 or SAE 50 automobile crankcase oil at
about 100°F has a viscosity of about 1000 SSU.)
© 2008 Taylor & Francis Group, LLC
12
Pump Characteristics and Applications
100
Head
CH
–0.6 × QNW
–0.8 × QNW
–1.0 × QNW
–1.2 × QNW
60
100
Rate of flow and efficiency
Correction Factors
80
CQ
80
60
Cη
40
20
00
33
00
22 0
6
17
20
13
0
88
0
66
0
44
0
33
0
22
6
17
2
13
88
65
43
32
10
20
15
Centistokes
00
,0
15 0
0
,0
10 00
80
00
60
00
40 0
0
30
00
20
00
15
00
10
0
80
0
60
0
40
0
30
0
0
20
0
2
15
1
10
80
Viscosity-SSU
60
600
400
300
200
150
100
80
60
40
30
20
15
40
Head in feet (per stage)
600
400
300
200
150
100
80
60
40
4
6
8 10
15 20
Rate of flow in 100 GPM at BEP
40
60
80 100
FIGURE 3.2
Performance correction chart for pumps handling viscous liquids. (Courtesy of Hydraulic
Institute, Parsippany, NJ; www.pumps.org.)
© 2008 Taylor & Francis Group, LLC
13
Special Hydraulic Considerations
SOLUTION
Enter the Figure 3.2 chart with 750 gpm, go up to 100 ft head, over to 1000
SSU, and then up to the preliminary correction factors.
CQ = 0.95
CH = 0.92 (for 1.0 × Q NW )
Cη = 0.635
Note that these preliminary correction factors allow a rough estimate
of the effect on performance of the viscous liquid described in this example. In this case, pumping this liquid will reduce capacity by about 5%,
total head by about 8%, and efficiency by about 36.5%, compared with
the pump’s performance when handling water.
Using the above preliminary correction factors, the equivalent water
rating is
Q W = 750 / 0.95 = 790 gpm
H W = 100 / 0.92 = 109 ft
60
200 13"φ
Total head
180
160
65
7'
8'
70
75
77'
10'
12'
16'
12"
77
140 11"
50
70
80
9"
30
60
20
40
20
0
200
0
50
1780 rpm
400
100
600
800
150
HP
HP
60
HP
HP
1000
30
10
1200
250
40
20
HP
Capacity
200
40
50
60
20'
75
120 10"
100
Meters
NPSHr
Model 3996 MT
Size 4 × 6–13
101–500
Impeller diameter 101–499
Pattern 54611
54612
Eye area 28 in2
Maximum allowable
HP = 60
Steel
Feet
Using this equivalent water rating, the pump whose curve is shown
in Figure 3.3 is chosen for the service, with an impeller diameter of 10
3/4 in. For this pump, the best efficiency point on the water curve at the
1400
300
1600 gpm
0
350 m3/h
FIGURE 3.3
Pump chosen for viscosity problem, Example 3.1. (Courtesy of Goulds Pumps, Inc., a subsidiary of ITT Corporation.)
© 2008 Taylor & Francis Group, LLC
14
Pump Characteristics and Applications
selected impeller diameter (QNW) is 850 gpm and 106 ft. Now re-enter
the correction chart Figure 3.2 with this BEP data from the water curve.
Enter the chart at 850 gpm, go up to 106 ft, over to 1000 SSU, and then up
to the revised correction factors.
CQ = 0.96
CH = 0.96 (for 0.6 × Q NW )
CH = 0.95 (for 0.8 × Q NW )
CH = 0.93 (for 1.0 × Q NW )
CH = 0.90 (for 1.2 × Q NW )
C η = 0.65
Now, on the water curve (Figure 3.3 with 10 3/4 in. impeller diameter),
pick off head, and efficiency at the following flows:
0.6 × Q NW = 0.6 × 850 = 510 gpm
0.8 × Q NW = 0.8 × 850 = 680 gpm
1.0 × Q NW = 1.0 × 850 = 850 gpm
1.2 × Q NW = 1.2 × 850 = 1020 gpm
Using these data points from the water curve and the revised correction factors from above, Table 3.3 can be constructed to generate four
points on the viscous pump curve. Water curve data points are labeled
QW, HW, and ηW. Viscous curve data points, labeled QV, HV, and ηV, are
obtained by multiplying the water data points by their respective correction factors.
Finally, the viscous brake horsepower is calculated for each of the
above four data points, using the formula for horsepower, Equation 2.16;
the given specific gravity, 0.9, and the values for QV, HV, and ηV from
Table 3.3.
TABLE 3.3
Viscous Pump Performance: Example 3.1
QW
HW
ηW
CQ
CH
Cη
QV
HV
ηV
0.6 × QNW
0.8 × QNW
1.0 × QNW
1.2 × QNW
510.0
125.0
67.5
0.96
0.96
0.65
490.0
120.0
43.9
680.0
117.0
74.2
0.96
0.95
0.65
653.0
111.0
48.2
850.0
106.0
77.0
0.96
0.93
0.65
816.0
99.0
50.1
1020.0
93.0
75.0
0.96
0.90
0.65
979.0
83.7
49.0
© 2008 Taylor & Francis Group, LLC
15
Special Hydraulic Considerations
1.00
C
H
.90
Correction factors
.80
.70
C
Q
.60
.50
C
η
.40
.30
.20
.10
0
220
0
176
0
132
880
660
440
330
43
88
66
32
20
15
10
7.4
4.3
Centistokes
220
176
132
.0
00
10,0 8000
0
40
0
30
0
20
0
15
0
0
1 0
8
60
0
600
0
400
300
40
30
20
15
10
8
6
Viscosity-SSU
10
15
0
Head in feet (per stage)
0
40
0
30
0
20 0
15
0
100
8
60
40
30
20
15
10
8
6
0
200
0
70
150
60
0
100
800
600
300
50
400
200
40
150
30
100
80
50
25
60
40
20
Rate of flow-gallons per minute (at BEP)
80 90 100
FIGURE 3.4
Performance correction chart for small pumps (capacities of 100 gpm or less) handling viscous
liquids. (Courtesy of the Hydraulic Institute, Parsippany, NJ; www.pumps.org.)
© 2008 Taylor & Francis Group, LLC
16
Pump Characteristics and Applications
BHPV = (Q V × H V × SG)/(3960 × ηV )
= 30.4 (for 0.6 × Q NW )
= 34.2 (for 0.8 × Q NW )
= 36.6 (for 1.0 × Q NW )
= 38.0 (for 1.2 × Q NW )
With the viscous performance values of QV, HV, and ηV from Table 3.3
and the above values for BHPV, a complete viscous performance curve
can now be drawn.
Note that the preceding discussion and example made no mention of
NPSH. Very little information is available to predict the effect of viscosity
on NPSHr. For values of viscosity below 2000 SSU, it is safe to use the NPSHr
from a curve based on water. For values above 2000 SSU, experimentation
or other experience with similar applications are about the only choices for
determining the suitability of the application. The best approach is to have
the largest margin possible between NPSHa and NPSHr.
The correction chart shown in Figure 3.2 can be used to very quickly decide
what will be the penalty in efficiency for pumping a viscous liquid with a
centrifugal pump. This is a quick way to decide if it makes sense to consider
a centrifugal pump for the application, or whether a positive displacement
pump should be used for the application. Remember that it may be possible
to heat the liquid to reduce its viscosity and thus make a centrifugal pump
a more viable option. Also, the computer software discussed in Section III
to follow can be used to correct centrifugal pump performance for viscous
liquids.
Figure 3.4 shows a correction chart similar to the one shown in Figure 3.2,
but for small pumps only (capacities less than 100 gpm). The Figure 3.4 chart
is used in the same manner as the example above, except that for smaller
pumps there is only one head correction factor CH.
III. Software to Size Pumps and Systems
A. General
Chapter 2 described manual techniques for sizing the components of a
pumping system, calculating pump TH, and then selecting a pump for this
service. The chapter explained how to develop a system head curve to study
how a pump will perform when pumping through various branches of the
piping system. Finally, Chapter 2 explained how to account for the use of
multiple pumps in the system. A thorough review of Chapter 2 reveals that
© 2008 Taylor & Francis Group, LLC
Special Hydraulic Considerations
17
the sizing of pumps and systems can be tedious and time consuming. This
is particularly true if the system contains multiple branches or loops, or if
multiple pumps are used in the system. Furthermore, if the pumped liquid is
something other than water, tables to determine friction losses through the
pipes, valves, and fittings may not be readily available. The same dilemma
might exist if the pipe wall thickness and material is different from standard
schedule 40 steel.
Because of these limitations, many pump system designers must make
their best approximation of pump head in these situations. This usually
results in the pump being oversized, and the adjustments to make the pump
operate at the desired capacity are done with control valves or orifices after
the system has been built and placed in operation.
As the example in Chapter 2, Section XII showed, oversizing a pump usually results in a more expensive pump, and a sizable increase in operating
cost. Maintenance costs may be higher too, because of excessive throttling
that may be required, or because the pump is operating at a nonideal point
on its performance curve. Furthermore, because of the tedious and repetitive calculations that must be performed by the engineer in designing a piping system, few systems ever actually are optimized when they are being
designed.
Fortunately, there are solutions to these dilemmas. Several computer
software packages are available that significantly reduce the time it takes
to design a piping system, while improving the capability to optimize the
system design. The software packages vary widely in their ease of use, their
capability to model complex systems, their flexibility to handle a range of
liquids and pipe materials, their user interface features, and their capability
to optimize the system design, as well as their cost.
B. Value of Piping Design Software
Piping software provides a simulated view of how the entire piping system
operates. An engineer can use this information to both design and optimize the piping system. A complete piping system model can be used at the
operating plant to provide everyone with a clear picture of how the system
operates. How the piping system model is used depends on the duties and
responsibilities of the user:
• Design engineers can use the information to size and optimize individual pipe lines. They are interested in calculating the pressure
drop or head loss in the pipe lines, and in optimizing the design.
• Rotating equipment specialists are interested in calculating the
design point needed for pump selection. After selecting the pump,
they want to see how it operates in the piping system.
© 2008 Taylor & Francis Group, LLC
18
Pump Characteristics and Applications
• Instrumentation and control engineers can use the results of a piping system model to determine the sizing requirements for control
valves.
• Plant or project engineers are interested in seeing how a piping system operates, as well as in maintaining proper system operation as
system changes are made.
• Plant maintenance personnel use the model to identify, isolate, and
correct problems in the piping system that cause excessive wear to
pumps and control valves.
• Finally, plant operators are interested in seeing how the system operates under a variety of expected operating conditions.
Ultimately, an accurate piping system model is a valuable resource for everyone involved with the system, and can provide a clear picture for the life of
the plant.
C. Evaluating Fluid Flow Software
When evaluating fluid piping software, it is important to determine what
features are necessary to meet the needs of the people who will use it, and
then choose the program that meets the majority of these requirements.
A variety of piping system software is currently available. To be effective, any piping software being considered should incorporate the following
features:
• Provide an easy-to-understand schematic showing how the pipe
lines, pumps, tanks, and components are connected.
• Perform the hydraulic network analysis needed to calculate the balanced flow rates and pressures for each pipe line in the system.
• Provide the results in a format that is easy to understand, and can
also be shared by everyone involved with the design, operation, or
maintenance of the system.
The remainder of this section describes the features found in a variety
of fluid flow programs. Each section describes the basic features needed to
meet minimum requirements, as well as “enhanced” features that provide
extra value to the user. All of the programs listed in Table 3.4 meet all of
the basic requirements, though all do not meet every one of the enhanced
features. Because software developers are continually adding new features,
no effort has been made to identify what enhanced features are available for
each program, and that is left to the person evaluating the software.
Because of the differences in features and ease of use of the available programs, interested readers are urged to study the product literature and demonstration programs available from the software developers.
© 2008 Taylor & Francis Group, LLC
19
Special Hydraulic Considerations
TABLE 3.4
Software for Designing or Analyzing Piping Systems and Components
Program
Manufacturer
Website
Phone Number
Design Flow
Solutions
Fathom
Flow of Fluids
Premium
FluidFlow 3
ABZ Inc.
www.abzinc.com
(800)747-7401
Applied Flow Technology
Crane Valve Co.
www.aft.com
www.flowoffluids.com
(800)589-4943
(888)206-2779
Flite Software N.I. Ltd.
www.fluidflowinfo.com
KYPIPE 2012
KYPIPE LCC Software
Center
Engineered Software, Inc
www.kypipe.com
44 28 71279227
(United Kingdom)
888-711-3051
(USA and
Canada)
(859)263-2234
www.eng-software.com
(800)786-8545
Sunrise Systems Ltd.
Epcon International
www.sunrise-sys.com
www.epcon.com
(800)367-3585
PIPE-FLO
Professional
PIPENET
SINET
D. Building the System Model
The first step in modeling a fluid piping system is to create the system model.
Because building the system model is the most time-consuming process, it is
important to select a program that streamlines this task as much as possible.
The ability to build the piping system model using a piping schematic drawing is a basic feature of any fluid piping program. This includes the ability
to place piping system items such as pumps, components, heat exchangers, tanks, and controls on the piping schematic. The various system items
are connected with individual pipe lines. Not only does the piping drawing show all of the items in the system, but it also provides the connectivity
information needed by the software to set up the equations needed to perform the balanced flow rate and pressure calculations.
Another basic requirement is the ability of the program to automatically
trace the loops in the system, set up the equations, and supply the initial
guesses needed to perform the calculations.
The user can build the piping schematic by selecting items from menus or
button bars, placing the items on the piping schematic, and entering information about each item. As the various items are added to the piping schematic, the user has visual indication as to what is added and what remains
to be added.
The data needed for the fluid flow calculations are entered into the program via dialog boxes or property grids. If the program relies on manual
entry of data, for example, an inside pipe diameter, the user must look up
© 2008 Taylor & Francis Group, LLC
20
Pump Characteristics and Applications
the information in a handbook and type in the value. If the program uses
pipe data tables to look up pipe properties, the user selects the pipe material, a corresponding pipe schedule or wall thickness from a displayed list,
and then chooses an available pipe nominal size. The program then looks
up the inside diameter from the pipe table. The program should also be
able to look up the fluid properties (density, viscosity, and vapor pressure)
at the operating temperature from data tables that are imbedded in the
software.
Following are the enhanced features found in some of the programs listed
in Table 3.4.
1. Copy Command
The ability to copy and paste items on the piping schematic simplifies the
creation of a piping system model. For example, when modeling a discharge
manifold, the ability to copy pipe segments and paste them into a manifold
saves time in building the model. The copy feature should copy both the
object on the piping drawing and the underlying design information. Some
programs support a group copy command, capable of copying multiple
objects at one time and pasting them to the piping schematic.
2. Customize Symbols
Each object placed on the piping schematic typically has a unique symbol shape.
The user can easily identify the pumps, tanks, and controls just by looking at the
drawing. If the program supports multiple symbol shapes for each type of item,
users can further customize the look of their piping schematic. Some of the programs listed even allow users to create their own customized symbol shapes.
3. CAD Drawing Features
The ability to zoom and pan around the piping schematic makes it easy to get
around large piping system models. The ability to insert new pipelines and piping system objects into the existing piping schematic greatly streamlines the
building of the model. The ability to move items around the screen and add new
vertices to pipelines lets the user give the piping schematic the look and feel
of an established drawing. The ability to place notes on the piping schematic,
change the size of the symbols and text, and change the pipe colors or line thickness on the schematic increases the presentation value of the program.
4. Naming Items
Each item in a piping system typically has a name or equipment identifier.
Many of the programs listed allow the user to attach a name to the object
and to display that name on the piping schematic. Using the user company’s
© 2008 Taylor & Francis Group, LLC
Special Hydraulic Considerations
21
naming convention on the piping system model makes it easier for everyone
to understand the model.
5. Displaying Results
The ability to display the calculated results on the piping schematic makes it
easier to understand the operation of the piping system. Seeing the flow rate
printed on the piping schematic next to the pipeline, or the total head of a
pump displayed next to the pump, provides the user with a clear picture of
how the piping system operates.
6. The Look of the Piping Schematic
If the piping schematic has the look of a generally recognized drawing, it is
easier for everyone to recognize the system model. Many of the programs
listed allow the user to create a piping schematic that looks like a standard
process flow diagram or piping and instrumentation (P&ID drawing). Often,
piping drawings are laid out in isometric view. This type of drawing provides an indication of elevation on the various items on the drawing. If the
program has an isometric grid feature, it is very easy to develop a professional-looking piping isometric.
E. Calculating the System Operation
The primary function of the piping system software is to calculate the balanced flow rates and pressures for the entire system. The program must
also factor in the effect of pumps, tanks, components, and control valves.
Reviewing the calculated results, the user can see how everything works
together as a system.
Hydraulic network analysis techniques are used to calculate the balanced
flow rates and pressures. Without getting into the technical details, the programs perform the calculations in the following ways:
• Determine how all the pipe lines in the system are connected and
then set up the necessary simultaneous equations needed to perform the balancing calculations.
• Establish an initial guess to balance the flow rates in the system.
• The initial guess is used to calculate the pressure drops around the
various loops in the system.
• If the loop pressure drops are not balanced, the program improves on
the flow rate guess, and then repeats the pressure drop calculations.
• Once the flow rates and pressure drop calculations have converged
to a solution, the results are displayed.
© 2008 Taylor & Francis Group, LLC
22
Pump Characteristics and Applications
All programs listed in Table 3.4 calculate the balanced flow rates and pressures as outlined above. A list of enhanced calculation features that provide
additional value to the program is included below.
1. Sizing Pipe Lines
If the user is working on a new piping system or adding a pipe to an existing
system, it is often helpful to size a single pipe line and determine its pressure drop based on sizing criteria (average fluid velocity, head loss per unit
length, pressure drop per unit length). This allows the user to optimize the
pipe line prior to adding it to the piping system model, by comparing the
pumping energy cost and initial installed cost for alternate pipe sizes. Many
of the programs listed have a pipe line-sizing feature built in, or have a separate pipe line-sizing program.
2. Calculating Speed
The speed at which the program calculates the balanced flow rates and
pressures is based on the programming skill of the company supplying the
software, the speed of the computer running the program, the efficiency of
the calculation method used in the program, and the speed of the programming language used in writing the program. For small piping systems of 20
or 30 pipe lines, the listed programs all have acceptable calculation speeds.
When working with larger piping system models, say 100 pipe lines or more,
the program calculation speed can become an important factor in software
selection.
3. Showing Problem Areas
The ability to highlight results in the piping model that are outside the
expected ranges provides the user with the ability to quickly identify and
correct problem areas in the system. For example, if the program provides a
list of pipe lines with high fluid velocities, the user can see which pipe lines
are undersized for the required flow rate.
4. Equipment Selection
Often, the calculated results derived by the program are used for equipment
selection. Knowing the TH and NPSHa is a requirement for proper pump
selection. The ability to select pumps and control valves from manufacturers’ catalogs is a valuable program feature. Once the equipment is selected,
the user should be able to easily add the pumps and control valves to the
model to provide a picture of how the selected equipment operates in the
system.
© 2008 Taylor & Francis Group, LLC
Special Hydraulic Considerations
23
5. Alternate System Operational Modes
A system is designed to meet specific flow rate and pressure requirements.
The original design values are also used to size pipe lines, pumps, and control valves. During normal plant operations, the piping system may run at
different conditions: lower flow rates, different tank levels, etc. The ability to
see how the system operates under these expected conditions provides an
invaluable record of the system design.
F. Communicating the Results
Sharing information with others is a vital part of any piping project. Once the
system is modeled, everyone in the project can share the results. This requires
the printing of the calculated results in a format that everyone can use.
Typically, hydraulic analysis software calculates the pressures and flow
rates in pipe lines and shows how pumps, components, and control valves
are operating. The software should provide all the information needed for
someone to determine how the system is operating. All the programs listed
in Table 3.4 have the ability to send the calculated results to a printer.
Below is a list of enhanced results features that provide additional value
to the program.
1. Viewing Results within the Program
The calculated results are the primary output of the program. This information is used to select equipment, optimize pipe lines, balance the system, or
identify when the results are outside the expected range. The ability to view
the results within the program without sending them to a printer not only
saves paper, but lets the user quickly try design alternatives.
2. Incorporating User-Defined Limits
When looking at the calculated results, it is often helpful to compare the
results to established design guidelines. For example, if the fluid velocity in
a pipe line is too great, that can be an indication that the pipe diameter is too
small for the required flow rate. The ability to set design guides or limits and
have the program indicate when these limits are exceeded makes it easier to
identify and correct problem areas within the system.
3. Selecting the Results to Display
The program’s printed results document how the system operates, and for
large systems, the reports can be lengthy. Some of the listed programs give
the user the option of choosing the reports the user wishes to print. Some
programs give the user the opportunity to choose only a part of the results
to print, thereby minimizing the size of the printed report. Finally, some
© 2008 Taylor & Francis Group, LLC
24
Pump Characteristics and Applications
programs let the user create their own customized reports by selecting the
fields to print, as well as the order in which they are listed.
4. Plotting the Piping Schematic
The piping schematic drawings for large systems can be very complex and
require the ability to send the report to a large format printer or plotter. This
is an important feature to consider if the user typically works on large piping systems. Some of the programs support tiled printing, the ability to print
the piping schematic on multiple pages. With tiled printing, the piping schematic is divided into tiles, with the portion of the system on each tile printed
to a single page. Once all the tiles are printed, they can be taped together,
providing a large format drawing.
5. Exporting the Results
The ability to export the design data and calculated results to a text file
makes it easy to share results between programs. For example, the results
can be imported into a spreadsheet program to provide input information
for external calculations or to allow for the generation of a custom report
using the page layout capability of the spreadsheet.
6. Sharing Results with Others
The ability to send piping system reports as e-mail attachments makes it easier to share results with others. Some programs create reports in a universal
format such as hypertext markup language (HTML) or portable document
format (PDF). These report files then become attachments to an e-mail message. The recipients then open the attached file with the appropriate reader
program, where they can view and print the results. This advanced printing
feature helps streamline the workflow.
7. Sharing Results Using a Viewer Program
To provide a greater degree of information sharing, some of the programs
listed in Table 3.4 support proprietary viewer formats that contain all the
calculated results and design information. The creator of the piping system
model sends a viewer file that the recipient opens with a corresponding
viewer program. The recipient can then view and print all the calculated
reports with the viewer program.
G. Conclusion
Anyone who designs or analyzes pump systems or selects pumps should
consider using computer software to assist in this effort. The capability of
© 2008 Taylor & Francis Group, LLC
Special Hydraulic Considerations
25
software solutions to save time, to reduce overall system costs, to handle
extremely complex systems, and to permit optimization of the system are all
important reasons to use these valuable tools.
In the past, fluid flow software was primarily used as an engineering and
design tool. Because of improved ease of use, more people working with
piping systems are using piping software to simulate the operation of the
system. When searching for fluid piping software, it is important to consider
the primary use of the software and which package best meets the user’s
needs. Any program being considered should have the basic features outlined in this section. It is left to the reader to make the determination of what
enhanced features are important and which fluid piping program will best
meet the needs of the user company.
H. List of Software Vendors
Table 3.4 provides a list of fluid piping software, along with the company and
contact information. Most programs listed have a demonstration disk that
can be downloaded from their website.
IV. Piping Layout
Improper piping layout, particularly on the suction side of the pump, can
lead to several serious problems with pump operation, and to excessive
pump maintenance costs. This section provides some guidelines on piping
design and layout to help minimize operation and maintenance problems.
Suction piping should be as short and as straight as possible. Short suction
lines minimize friction losses to help ensure adequate NPSHa. Straight suction lines help the liquid enter the impeller in a straight line, to minimize
uneven loading of the impeller and bearings. Keep suction line velocities in
the 4- to 6-ft/s range. Lower suction velocities are permissible if the pumped
liquid is free of solids that must be kept in suspension.
Care should be taken if an elbow is located at the suction of a double
suction pump. Chapter 4, Sections II.B and VI, describes double suction
pumps. One of the primary benefits of a double suction pump is that, in
single-stage configuration, the impeller produces almost no axial thrust. (See
Chapter 4, Section II.D, for a discussion of axial thrust.) If an elbow is located
at the suction of a horizontal double suction pump, the elbow should be in a
vertical rather than horizontal orientation. Having an elbow in a horizontal
orientation permits an uneven division of the flow between the two sides
of the double suction impeller, which leads to excessive axial thrust. If the
elbow must absolutely be located in the horizontal plane, it should be located
several pipe diameters away from the pump. Also, several companies make
© 2008 Taylor & Francis Group, LLC
26
Pump Characteristics and Applications
flow straightening diffusers that can be located at the pump suction downstream of the elbow to help straighten the flow before it enters pump.
Check valves are notorious for having problems with sticking, getting
debris trapped in them, or not seating properly, so a check valve located at
the discharge of a pump should always be able to be isolated from the system
pressure for maintenance. This is illustrated in Figure 3.5.
Allowing air to get into a pump can reduce head and flow (as demonstrated in Figure 3.6), increase noise levels, and most importantly, lead to
increased radial bearing loads, which can cause premature failure of bearings and seals, or shaft breakage. Sometimes, this air entrainment is confused
with cavitation (Chapter 2, Section VI) because the symptoms are similar.
Check
valve
Eccentric
reducer
Long radius
elbow
Gate
valve
Suction pipe slopes
upwards from source
of supply
Correct
Foot valve (if used)
Strainer
Air pocket because eccentric reducer
is not used and because suction
pipe does not slope gradually
upward from supply
Gate
valve
Check
valve
Gate valve should not be
between check valve and pump
Wrong
FIGURE 3.5
Use an eccentric rather than concentric reducer, and slope suction lines upward from the
source if operating on a lift. Check valve should be able to be isolated for service. (Courtesy of
Goulds Pumps, Inc., a subsidiary of ITT Corporation.)
© 2008 Taylor & Francis Group, LLC
27
Total head
Special Hydraulic Considerations
2% Air
0% Air
4% Air
6% Air
Flow rate
FIGURE 3.6
Effect of air on the performance of a centrifugal pump.
Several aspects of pump system piping that can introduce air into a pump
are described below.
For a pump operating on a suction lift, the pressure inside the suction
line above the liquid level in the suction vessel is below atmospheric. Thus,
the flange connections in the suction line must be properly sealed with gaskets or O-rings to keep air from leaking into the suction line at the flange
joints.
Suction piping is generally one size larger than the suction flange of a
pump to minimize the friction losses in the suction piping. This means that
a reducer is required at the pump suction flange. The reducer should be an
eccentric, rather than a concentric reducer (Figure 3.5). If the reducer at the
pump suction flange is concentric, it would have a small cavity at the top of
the reducer next to the pump suction flange, which is higher than the suction
flange itself. This is a cavity where any air that has been introduced into the
suction line could accumulate and cause problems with pumping.
For systems with the pump operating on a suction lift, horizontal pipe
runs on the suction line should be continuously sloping upward from the
sump to the pump (Figure 3.5). The reason for this requirement is the same
as in the previous paragraph. To do otherwise might create a cavity located
higher than the top of the pump suction flange, where air might accumulate
and eventually pass into the pump or partially block flow into the pump.
Return lines back to a pump supply vessel should not be allowed to free
fall into the vessel (Figure 3.7), as the turbulence caused by the falling water
can produce air bubbles, which can move into the pump and accumulate.
This guideline applies whether the pump is operating on a suction lift or
on a suction head. Return lines should always be submerged below the liquid level, and located as far as practical in the vessel away from the pump
© 2008 Taylor & Francis Group, LLC
28
Pump Characteristics and Applications
Pump suction
Recommended
Recommended
Baffle
Pump
suction
FIGURE 3.7
Return lines back to a pump suction vessel should not be allowed to free-fall into the vessel.
(Courtesy of Goulds Pumps, Inc., a subsidiary of ITT Corporation.)
suction connection. An additional measure is to separate the return line
from the pump suction with a baffle, so that any air bubbles that are permitted to enter the system will float to the surface in the supply vessel, rather
than being reintroduced into the pump.
Vortex formation is another phenomenon that can introduce air into a centrifugal pump pumping from an open sump. A vortex is a rotating swirl
of water, often observed when the plug is pulled from a filled sink. A fully
developed vortex in a supply vessel can draw air down into the pump suction. This can occur if the liquid drops below a prescribed minimum level,
which is a function of the velocity of the liquid at the suction connection on
the supply vessel. The reference point from which to measure this minimum
submergence for vortex suppression with various suction configurations is
illustrated in Figure 3.8.
Although manufacturers are not in complete agreement as to the precise
details, Table 3.5 shows suggested values of minimum submergence (liquid
surface to top of suction pipe) to prevent vortices, as a function of velocity
in the suction pipe. Where there is a bell at the inlet (e.g., a vertical turbine
pump or the suction pipe for a pump on a suction lift), velocity is calculated
at the widest part of the bell, and submergence is measured from the liquid
surface to the bell lip.
In some installations, it is impossible or impractical to keep the liquid level
in the supply vessel above the recommended minimum level. In these situations, several steps can be taken to break up the vortex or keep it from entering the pump. One arrangement (Figure 3.9) uses a flat baffle plate located
just above the pump suction connection. This plate helps prevent the vortex from fully forming and entering the pump suction, because the vortex
would have to form in a horizontal plane, which is less likely to occur.
© 2008 Taylor & Francis Group, LLC
29
Special Hydraulic Considerations
Pump
suction
Minimum
submergence
Minimum
submergence
Minimum
submergence
Pump
suction
Pump
suction
FIGURE 3.8
Reference point from which to measure minimum submergence for various suction
configurations.
TABLE 3.5
Minimum Submergence for Vortex Suppression
Suction Velocity (ft/s)
2.0
4.0
6.0
8.0
10.0
Minimum Submergence (ft)
1.0
2.0
4.0
6.5
9.5
Another strategy, also shown in Figure 3.9, uses a cross-hatched baffle in
the suction piping to break up the rotational swirling motion of the vortex.
With vertical wet pit pumps, still another technique for breaking up vortices
uses the suction screen normally attached to the bell of the pump for the
purpose of preventing foreign materials above a certain size from entering
the pump. The screen serves the same function as the cross-hatched plate
just described, breaking up the vortex before it enters the pump.
Even in a system where the pump has a positive static suction head, it is
good practice for the suction line to be continuously rising from the suction
vessel connection until it reaches the pump. If there is a high point (above the
© 2008 Taylor & Francis Group, LLC
30
Pump Characteristics and Applications
Flat baffle
Water depth
Side view
Suction
pipe
Baffle
smooths
out vortex
Flat
baffle
Top view
FIGURE 3.9
A flat baffle plate located just above the pump suction connection or a cross-hatched baffle in
the suction piping, minimize the possibility of vortex formation. (Courtesy of Goulds Pumps,
Inc., a subsidiary of ITT Corporation.)
Low pressure region
for air pocket to form
Aerated
water
FIGURE 3.10
A loop in the suction line can cause a vacuum near top of the loop and can lead to an accumulation of air at the high point of the loop.
© 2008 Taylor & Francis Group, LLC
Special Hydraulic Considerations
31
pump inlet) anywhere in the suction line, this is the location where any air
or highly aerated water that enters the suction line (due to any of the reasons
described above) can accumulate and cause problems. For this same reason,
it is generally not recommended to have a suction line that dips below the
liquid level but comes out of the top of the suction vessel, or otherwise loops
up above the pump inlet (Figure 3.10).
V. Sump Design
A thorough treatment of this subject is beyond the scope of this book, but a
few words deserve to be written on the subject of sump design. Many installations require multiple pumps in a common sump or intake structure. This
is often the case for plant makeup water pumps, cooling water supply, and
fire pumps, to name some common applications. The pumps are usually horizontal split case, double suction pumps (see Chapter 4, Section VI) operating
on a lift, or vertical turbine pumps (see Chapter 4, Section XI).
The primary objectives for the designer of an intake sump with multiple
pumps are to ensure that each of the pumps in the sump is allowed to receive
its intended flow rate, to make certain that the liquid entering each pump is
moving in a straight line as it enters the pump, and to prevent the possibility of
vortex formation that could lead to air introduction into the pump and system.
The Hydraulic Institute (Ref. [4]) offers guidelines to help meet these design
objectives. These guidelines establish key dimensions for sumps containing
multiple pumps, such as the optimal distance the pumps should be located
from each other and from adjacent walls, acceptable and recommended
angles for pipe openings to open channels, etc. Suction velocity in an open
channel is recommended to be 1 ft/s or less and the flow should approach
the pump in as straight a line as possible. The Hydraulic Institute offers some
guidelines as to the configuration of multiple pumps. For example, multiple
pumps should not be lined up in a narrow sump in the direction of the main
flow path because of the likelihood that the final pump in the line would not
receive as much flow as the first pump in the line.
If Hydraulic Institute Standards are used as a design guide for designing
a multiple pump sump, a conservative design will result, one that will likely
cause no pumping problems. (When asked to comment on proposed sump or
intake structure designs, most pump manufacturers will advise the engineer
or owner to follow Hydraulic Institute design guides.) Although this approach
is conservative, it may result in a more expensive intake structure. One possible
alternative to explore early in the design phase for a plant intake system is a
sump model test. This test, using dimensionless parameters such as the Froude
number, can create a scale model of the sump and the profiles of the various
pumps to be located in the sump. Then the flow to the sump can be modeled
© 2008 Taylor & Francis Group, LLC
32
Pump Characteristics and Applications
to allow for physical observation of the flow distribution to various parts of the
sump and for observation of the formation of eddies and vortices. Also, the
sump model design and pump orientation can easily be altered in a model test,
by adding or moving baffles, rearranging the pumps in the sump, or changing
the liquid level in the sump, until the best arrangement and design is achieved
given the physical limitations of the sump. This model test approach can sometimes result in significant construction cost savings for an intake structure.
Sump model tests are available from some manufacturers of large intake
pumps, from civil or mechanical engineering departments at major universities, and from a number of independent testing laboratories.
VI. Field Testing
A. General
Field testing usually involves taking measurements of pump flow, total head,
and power consumption from an operating pump. Using these test data,
along with a knowledge of the motor efficiency, the pump efficiency can be
computed. Thus, the complete operating characteristics of the pump can be
produced at a range of flow values, allowing a complete new set of pump
curves to be generated in the field.
Some people view a field test as a way to verify whether or not they were
“cheated” by the pump supplier. This perceived benefit is often not realistically achievable for the simple reason that measurements taken in most field
tests are less accurate than the tests the manufacturer performs in the test
lab. Measuring flow in an installed pump is particularly difficult to do accurately without expensive flow measuring equipment.
A calculation of pump efficiency for an installed pump requires the measurement of flow, TH, and power draw to the motor (plus knowledge of the
motor’s efficiency at its operating load). When these three variables are measured to calculate efficiency, the accuracy of the result is equal to the square
root of the sum of the squares of the accuracies of the individual variables
tested. This is expressed by the following formula:
%E =
% 2Q + % 2H + % 2P
(3.4)
where %E is the accuracy of pump efficiency, %Q is the accuracy of flow measurement, %H is the accuracy of total head measurement, and %P is the accuracy of power measurement.
For example, suppose that pump TH and motor kilowatt draw can be
computed to within 2% in a given installation, but pump flow can only be
© 2008 Taylor & Francis Group, LLC
33
Special Hydraulic Considerations
measured to within 5%. In this case, the expected accuracy of the computed
efficiency, by Equation 3.4, is
%E =
52 + 2 2 + 2 2 =
33 = 5.7%
The preceding example illustrates that the accuracy of efficiency testing is
no more precise than the accuracy of the least accurate variable measured,
which in the majority of cases is flow. Very little benefit would be gained, in
the preceding example, by modification of the test assembly to allow more
accurate measurement of pump total head or motor kilowatt draw.
Despite its shortcomings, field testing of pumps does have several major
benefits. The primary benefit of field testing a pump is to establish a benchmark of total head, flow, and power or efficiency at one or more points of
operation on the pump curve when the pump is first installed in a new
system. These benchmark data points can then be used to compare against
pump performance in subsequent periodic field tests. These tests can be a
big help in calling attention to a developing problem before significant deterioration can occur.
One pump malfunction that can be observed through field testing is
operation of the pump at an unhealthy point on the H–Q curve (outside the
preferred operating range). Another measurable malfunction is cavitation,
causing the pump curve to drop off from where it should be at the point
where cavitation commences, as shown in Figure 2.13. Another problem
observable through a field test is a blocked or partially blocked suction line,
which allows the pump to hit its factory tested curve only at shutoff, and falls
below the new pump curve at all other points. Finally, a field test can show
when a pump is experiencing excessive recirculation leakage, signifying the
need to replace wear rings on a closed impeller and to reset clearances or
replace an open impeller. The excessive recirculation is indicated by curve
points falling low on head and capacity compared with what they should be
and by a reduction in efficiency and an increase in power. Early detection
of these problems allows the pump to be returned to a safe operating point,
ensures that system problems can be checked out, or permits better planning
of a pump repair to restore recirculation clearances.
B. Measuring Flow
The first component of pump field testing discussed here is flow measurement. There are quite a few different methods to measure pump flow. They
vary as to the accuracy of their measurement, their cost, and their complexity. Many are useful only for a specific range of flows or for certain types of
liquids. Most are located at a fixed place in the piping system, whereas a few
types are more portable. Most cause a pressure drop as liquid flows through
© 2008 Taylor & Francis Group, LLC
34
Pump Characteristics and Applications
the measuring device. Being exposed to the pumped liquid, many flowmeter
types are subject to a loss of accuracy over time due to the effects of corrosion and/or erosion. The following are brief discussions on the major types
of flow measurement systems.
1. Magnetic Flowmeter
Magnetic flowmeters are commonly used by pump manufacturers in their
laboratory testing facilities, and are also now often found in power plants and
other industrial applications. These instruments are very accurate and easily
adaptable to computer-based data acquisition instruments. Their shortcomings are that they are expensive, not portable, and usable only for electrically
conductive liquids. (They will not work on hydrocarbons, for example.) Also,
the electrical contacts that are exposed to the liquid in a magnetic flowmeter
can become contaminated, thus reducing accuracy.
2. Mass Flowmeter
Mass flowmeters work on the Coriolis principle. They are highly accurate
and work on a broad range of liquid types. Disadvantages include the fact
that they are expensive, are only available for fairly small pipe sizes (up to
about 8 in), and have a substantial pressure drop.
3. Nozzle
A nozzle is a special type of venturi (see Section VI.B.10 to follow), with a
converging section and a throat, but no diverging section. With a nozzle, the
throat empties to atmosphere. Flow is a function of the pressure immediately
upstream of the converging section, measured with a pressure gauge.
A nozzle has the same shortcomings as the venturi meter. Its accuracy
can deteriorate rapidly if deposits form on its interior surfaces or when wear
occurs. Additionally, because the nozzle must discharge to atmosphere, its
use is limited to very small systems or pumps that can discharge freely to
atmosphere.
4. Orifice Plate
The orifice plate is one of the most commonly used industrial flow measurement devices. Similar to a venturi (see Section VI.B.10 to follow), the flow
through an orifice plate is a function of the differential pressure across the
device. Its chief advantages are its low initial cost and ease of relocation in a
piping system. The main disadvantages of orifice plates include limited turndown ratio (range of flow rate that can be measured), high pressure drop,
and loss of accuracy as the edges of the orifice wear.
© 2008 Taylor & Francis Group, LLC
Special Hydraulic Considerations
35
5. Paddle Wheel
Practical only on quite small systems, paddle wheel flow instruments are
the most common devices used for residential water meters. These paddle
wheels simply count the rotations as flow goes past them, with the flow being
proportional to the number of rotations. They are inexpensive, but not very
accurate, and are usually only used on water at relatively low temperatures.
6. Pitot Tube
A pitot tube is a relatively simple flow measuring device and is especially
useful for large systems handling ambient temperature water. The pitot tube
in its simplest form consists of a tube with a right-angle bend, which, when
immersed with the bent part pointed directly into the flow, indicates flow
velocity by the distance water rises in the vertical stem. It is often installed
in a side opening of the pipe with packing to allow a traverse of the pipe to
establish a velocity profile. This allows the computation of an average velocity and the resulting flow.
7. Segmental Wedge
With this flow measuring device, flow is a function of the differential pressure measured across a wedge located in the pipe line. Although measuring
flow quite accurately, these devices are expensive and not portable.
8. Turbine Meter
A sophisticated version of a paddle wheel, this type of flowmeter works on
the positive displacement principle that flow is proportional to speed. Like
the paddle wheel, the turbine meter, acting like a gear pump in reverse,
measures rotational speed and correlates it to flow. These meters are highly
accurate, and are sometimes used as “custody transfer” meters on pipe line
systems.
Shortcomings of turbine meters include the fact that they are quite expensive, have a fairly high pressure drop, are subject to wear by abrasives in the
liquid, and are not portable. Many also have the disadvantage of having limited turndown ratio, so that if the application requires flow measurement over
a wide range of flows, it is possible that more than one meter will be required.
9. Ultrasonic Flowmeter
These flowmeters work with a pair of transducers mounted on the outside
of the pipe, which send and receive ultrasonic signals. Taking into account
the sonic velocity of the pipe material and the liquid, flow is proportional to
the lag between the signal going with the flow and the one going against it.
© 2008 Taylor & Francis Group, LLC
36
Pump Characteristics and Applications
The distinct advantage of ultrasonic flowmeters over most others types
of flow measuring devices is that the meter is nonintrusive. That is, it is
mounted on the outside of the pipe, and not inside the pipe. This means it is
not subject to a reduction in accuracy due to wear, it can be worked on without penetrating the pipe, and it is portable so that it can be used in numerous
locations. Accuracies to less than 1% are achievable with these devices. They
are also capable of high turndown ratios.
Shortcomings of ultrasonic flowmeters are that they are somewhat temperamental with liquids containing gases, and the flow is subject to being
distorted by upstream or downstream valves or fittings if they are closer
than about 10 pipe diameters. Also, the setup may have to be modified if different liquids are being measured at the same location.
10. Venturi
Venturis are commonly used for industrial flow measurement and may be
the most economical choice of flow instrument for water supply systems.
A venturi has a converging section, a throat section, and a diverging section. The flow through the venturi is a function of the upstream and throat
dimensions, and the differential pressure across the venturi.
The advantages of a venturi include that it requires very little maintenance
and causes very little head loss. The major shortcoming of this method of
measuring flow is the fact that while the venturi is quite accurate when
newly installed, its accuracy can deteriorate if deposits form on its interior
surfaces or when wear occurs.
11. Volumetric Measurement
Volumetric method of flow measurement is usually practical only for very
small systems and is only accurate if the pump flow rate is very stable. It consists simply of measuring how long a pump operating at steady state takes to
empty or fill a known volume in a vessel. The measured volume, divided by
the time, equals the average pump capacity.
12. Vortex Flowmeter
A vortex flowmeter has a protrusion in a pipe that causes a vortex. It has
been likened in principle to a flag fluttering in the wind, where the faster the
wind flows, the higher the frequency of the fluttering of the flag. Similarly,
the frequency of the vortex can be proportioned to flow.
These flowmeters are not recommended for applications with a Reynolds
number less than about 10,000. This means they have limited turndown
ratio, and also will not work at low flow velocities. Additionally, they lose
their accuracy as the protruding body is worn or coated with deposits.
© 2008 Taylor & Francis Group, LLC
37
Special Hydraulic Considerations
C. Measuring TH
The most commonly used method to measure total head is by means of pressure gauges (or vacuum gauges in the case of a vacuum in the suction). The
gauge setup is shown on Figure 3.11. The formula for TH when there is a
vacuum in the suction is
TH = discharge gauge reading (ft)
+ vacuum gauge reading (ft)
+ distance between point of attachment of vacuum gauge and
centerline of discharge gauge, h (ft)
 V2
V2 
+ d − s 
2g 
 2g
(3.5)
When there is a positive pressure in the suction, a pressure gauge is used
on the pump suction rather than a vacuum gauge, and the formula becomes
TH = discharge gauge reading (ft)
− suction gauge reading (ft)
+ distance between centerlines of
discharge and suction gauges, h (ft)
 V2
V2 
+ d − s 
2g 
 2g
(3.6)
The third term in each of these two formulae is positive, assuming that
the discharge pressure gauge is located higher than the pressure or vacuum
h
h
Vacuum
Pressure
FIGURE 3.11
Measuring TH using pressure or vacuum gauges. (Courtesy of Goulds Pumps, Inc., a subsidiary of ITT Corporation.)
© 2008 Taylor & Francis Group, LLC
38
Pump Characteristics and Applications
gauge on the suction. If not, the term is negative. Actually, the two gauges are
often so close in elevation to each other that this term is ignored.
The final term in the above two equations is the change in velocity head
that is caused by the difference in size of the suction and discharge lines at the
gauge locations. It is found by first measuring flow (see Section VI.B above) or
approximating it if flow is not being measured. Then V2/2g is calculated at the
suction and discharge gauge locations or is looked up in friction tables such
as Table 2.1. As discussed in Chapter 2, this difference in velocity head is often
small enough compared with the total pump head that it can safely be ignored.
Note that when using pressure gauges, distances are measured from the
centerline of the gauge. With vacuum gauges, distances are measured from
the point of attachment of the gauge to the pipe.
Make sure that gauges used to test pumps are in calibration and that they
are the proper pressure rating to get an accurate reading at the pressure levels expected in the pump system.
Differential pressure transmitters can also be used to measure pump head,
converting differential pressure in psi to a 4 to 20-mA signal. These are substantially more expensive than gauges, but can be used to continuously monitor pump head, and are more accurate than most gauges.
A manometer (Figure 3.12) can also be used to measure low pressure or
vacuum. Manometers can be easily field fabricated using a liquid of a known
specific gravity. Water and mercury are common liquids used in manometers.
Pipe
line
Pipe
line
Vacuum
Pressurized
hpress
Water
hvac
Mercury
FIGURE 3.12
Manometers can be used to measure vacuum or low pressure.
© 2008 Taylor & Francis Group, LLC
Special Hydraulic Considerations
39
D. Measuring Power
The power required by a pump and motor (input power) can be measured
by a kilowatt transducer if one is available. More commonly, to obtain input
kilowatts, volts, and amperes are measured using a voltmeter and an ammeter (for three-phase power, one uses the average of the amperage measured
in each of the three legs). The following formula yields input KW:
KWin =
1.732 × voltage × ampere × power factor
1000
(3.7)
The constant 1.732 (square root of 3) in Equation 3.7 is used for three-phase
power measurements. For single-phase, that term in Equation 3.7 becomes
1.0.
The power factor in Equation 3.7 can be obtained from the motor manufacturer and varies with motor load.
Note that a power reading in KW can be converted to horsepower (HP) by
the following equation:
HP = KW × 1.34
(3.8)
E. Measuring NPSH
Chapter 2, Section VI.E, described the test done by the manufacturer at the
factory to produce the pump’s NPSHr curve. This test can also be carried
out in the field but, practically speaking, it seldom is, due to the complexity of the test setup and the difficulty in accurately measuring the required
parameters.
The NPSHa for a system at the suction of an operating pump can be measured in the field by means of a gauge pressure reading taken at the pump
suction, using the following formula (all terms are expressed in feet):
NPSHa = Pb ± G + Hv – Hvp
(3.9)
where Pb is the barometric pressure (ft); G is the gauge reading at the pump
suction (ft) (plus if above atmospheric, minus if below atmospheric), corrected, if necessary, to the pump centerline; Hv is the velocity head in the
suction pipe at the gauge connection (ft); and Hvp is the vapor pressure (ft),
based upon measured temperature.
NPSHa can be measured by the above method for a system where the
pump is suspected of cavitating. The measured value of NPSHa can then be
compared with the value of NPSHr shown on the pump curve.
© 2008 Taylor & Francis Group, LLC
4
Centrifugal Pump Types and Applications
I. Overview
Now that the reader has been introduced to pumps in Chapter 1 and has
developed a solid foundation in hydraulic theory related to centrifugal
pumps from Chapters 2 and 3, it is time to begin looking at hardware. This
chapter devotes itself to a discussion of the mechanical features and applications of the most important types of centrifugal pumps.
The chapter begins with a discussion on impeller types, because the impeller is the most important component of the pump. The chapter describes the
two most common types of impellers (open and closed), with a discussion on
the physical features, types of leakage joint, and applications most suited to
each type of impeller. Single and double suction impellers are described. The
explanations of these major impeller types lead into discussions of impeller
wear rings, suction specific speed, axial thrust and thrust balancing, filing of
impeller vane tips, and solids handling impellers.
Following the introduction to impellers, the bulk of the remainder of
Chapter 4 examines the most common types or configurations of centrifugal
pumps. Each pump type is described as to its physical configuration and to
its important hydraulic and mechanical design features. The applications for
which each pump type is best suited are described. Note that discussions on
sealing systems are presented separately in Chapter 5.
The reader is again referred to Appendix A at the end of this book, which
lists the major suppliers of centrifugal and positive displacement pumps
in the United States, along with an indication of the types of pumps they
manufacture. As mentioned in Chapter 1, Appendix A should by no means
be considered all-inclusive, but merely a listing of the manufacturers with
whom the author is familiar. The reader should be able to use an Internet
search, a Thomas Register, or similar directory to locate particular manufacturers shown in Appendix A. For any pump application being considered,
the best advice is to obtain information from a number of the manufacturers
for the types of pumps being considered.
One of the early decisions that must be made in centrifugal pump selection is the type of pump configuration that will be used. The solution is not
41
42
Pump Characteristics and Applications
always obvious, and often there are a number of alternatives. Consider, for
example, a very common pump application problem, that of emptying a
liquid out of a sump where the liquid level in the sump is below the level
where the pump would normally be located. There are at least a half dozen
different centrifugal alternatives for this application problem. The choices
could include a submersible pump, a pump with only the impeller and casing immersed in the liquid (either a vertical column sump pump or a vertical
turbine pump), a self-priming centrifugal pump, a non-self-priming end suction pump with a priming system, or with a foot valve, or an ejector system.
Any one of these choices might work for the application, but the question is,
which one is the best, given the particular conditions? Which one results in
the lowest first cost for the pump, which has the lowest installed cost (not
the same as first cost of the pump alone), which uses less energy to operate, and which has the lowest expected maintenance costs? The answers to
these questions depend, among other factors, on the flow and head required,
physical limitations of the installation, the corrosive nature of the pumped
liquid, whether there are abrasives or solids in the liquid, the depth of the
sump, and the liquid temperature. By a careful examination of the most common pump types, this chapter provides many of the tools to help make this
type of application decision.
Chapter 4 includes a discussion of pump standards and design specifications, with particular emphasis on three important standards for centrifugal pumps: (1) American National Standards Institute (ANSI), (2) American
Petroleum Institute (API), and (3) International Standards Organization
(ISO).
Finally, Chapter 4 concludes with sections on two of the most important
pump accessories: (1) couplings and (2) electric motors.
II. Impellers
A. Open versus Closed Impellers
There are two major classes of mechanical designs for centrifugal pump
impellers: open and closed. These two types of impellers are illustrated in
Figure 4.1. The distinction between these two impeller types has nothing to
do with the impeller specific speed, head/flow characteristics, or number and
shape of impeller vanes. Rather, the distinction has to do principally with
the way the leakage joint for the impeller is designed. The leakage joint exists
because the pressure on the discharge of the impeller is higher than the pressure on the suction (or front side) of the impeller. Because the impeller is rotating
within a stationary casing, there must naturally be a certain amount of running clearance between the impeller and casing. Therefore, a certain amount of
43
Centrifugal Pump Types and Applications
Open impeller
Closed impeller
FIGURE 4.1
Open and closed impellers. (Courtesy of Goulds Pumps, Inc., a subsidiary of ITT Corporation.)
liquid tends to leak from the discharge side of the impeller back to the suction
side, through the leakage joint. It is desirable to minimize the amount of leakage across the leakage joint in a centrifugal pump. Recall from Chapter 2 that
volumetric loss is one of the contributing factors to the inefficiency of a pump.
Figure 4.1 shows both an open and a closed impeller, both viewed from
the suction side. The primary difference is that with the open impeller, the
impeller vanes are clearly visible when viewed from the suction side of
the impeller. The closed impeller, on the other hand, has a shroud covering
the vanes on the suction or front side, and an axially oriented hub that provides the inlet for the liquid into the vane passageways.
The easiest way to understand how the leakage joints for the two types of
impellers work is to look at images of pumps with the two types of impellers. Figure 4.40 shows a pump with an open impeller. The leakage joint for
the open impeller consists of the front edges of the impeller vanes, which are
machined to a contour that is identical to the contour of the casing adjacent
to the impeller. To minimize leakage across the leakage joint, the front edges
of the impeller vanes must be kept very close to the mating face of the casing, while at the same time avoiding the possibility of the rotating impeller
rubbing against the casing. A common axial clearance or impeller setting to
achieve this is 0.012 to 0.015 in.
Open impeller pumps are generally considered a good choice for pumping
liquids that contain solids or stringy materials. The exposed vanes reduce
the likelihood of debris becoming trapped as it passes through the impeller.
Most open impeller pumps have the capability to adjust the impeller axially
in the field to account for wear on the front edges of the impeller vanes. This
wear can occur as the liquid squirts past the leakage joint from discharge
to suction, and is exacerbated if the liquid contains abrasives or is corrosive. For open impellers with pump-out vanes (see Section II.D below), adjustment of the axial clearance affects the resulting pressure in the stuffing box
or seal housing, and the pump axial thrust, in addition to pump efficiency.
44
Pump Characteristics and Applications
(See Chapter 5, Sections III and IV, for a more complete discussion of stuffing
boxes and seal housings.)
Pump manufacturers use different methods to adjust open impellers. The
pump in Figure 4.40 carries out this axial adjustment by means of two sets
of machine bolts at the coupling end of the pump, located on a common
bolt circle. Half of the bolts are fitted into drilled and tapped openings in
the bearing housing, while the other half act as jacking screws. When it is
desired to adjust the impeller axial setting, the pump is shut down, and the
bolts that go into the tapped openings are tightened, moving the entire rotating assembly toward the suction end of the casing. At the same time, the
technician turns the pump by hand from the coupling end. When the front
edges of the impeller vanes come in contact with the casing, it begins to rub,
and the impeller axial setting is then essentially zero. The impeller is then
adjusted away from the casing by turning the other set of machine bolts (the
ones that act like jacking screws). This moves the whole rotating assembly
away from the inlet side of the pump casing.
The amount of axial movement of the impeller during this adjustment is
measured in one of two ways. One method requires the gap behind the plate
where the bolts are located to be measured using feeler gauges, and then
remeasured as the jacking screws widen this gap, until the desired axial setting is achieved. Another method uses a dial indicator attached to the pump
base, with the pin of the indicator on the plate where the jacking screws are
located. Either method allows the impeller to be located a precise distance
from the front end of the pump casing.
A major benefit of this style of impeller is that its leakage rate can be reduced
by resetting the impeller in the field as just described, without disassembling
the pump. This tightening of the impeller clearance improves the pump’s
efficiency, and increases flow and/or head. However, this readjustment of
clearance as the impeller wears can only be done a limited number of times
without affecting pump performance. As the front edges of the impeller
vanes wear due to the corrosive/erosive effect of the liquid squirting past the
leakage joint, the width of the impeller vanes is actually being made smaller.
Eventually, if the vanes wear enough, the reduced impeller width causes the
capacity of the pump to begin to diminish. At that point, the only way to
bring the pump back to its original hydraulic performance is to replace the
impeller entirely. Also, as discussed in Section II.D below, axial adjustment
of this impeller type affects stuffing box pressure and axial thrust.
Another open impeller type has no pump-out vanes and has an axially
oriented (parallel to the shaft) leakage joint on the stuffing box side of the
impeller. This design minimizes the effect of axial adjustment but gives up
some of the solids handling and wear capability of the open impeller.
Both types of open impellers require a locked thrust bearing that limits
axial movement to 0.001 in.
A very common type of closed impeller (also called enclosed) is illustrated in
Figures 4.9 and 4.11. With this type of impeller, the leakage joint is the axially
Centrifugal Pump Types and Applications
45
oriented annular space between the rotating hub on the suction side of the
impeller and the space next to the hub on the pump casing. Quite often, the
rotating impeller hub or the mating part on the pump casing, or both, will be
fitted with wear rings. These are replaceable rings that, if wear occurs in the
leakage joint from corrosion and/or erosive wear, can be more economically
replaced than the alternative of replacing the entire impeller or casing. The
pump in Figure 4.11 is fitted with a wear ring in the pump casing.
Because of the axial orientation of the leakage joint for the closed impeller
described above, the axial setting on the closed impeller is not nearly as critical to the pump’s performance as is the case with the axial setting on an open
impeller. With a closed impeller, where the leakage joint hub may be nearly
an inch long (varying with impeller size), moving the impeller axially a tenth
of an inch or so in either direction has a very minor influence on the amount
of leakage across the leakage joint. The amount of leakage across the joint is
primarily a function of the differential pressure and the diametral clearance
at the impeller front hub. Therefore, the axial setting is primarily to ensure
that the impeller is not rubbing against the front side of the casing.
A shortcoming of this type of closed impeller construction is that, although
the leakage joint can be renewed on a pump fitted with wear rings by replacing the rings, the pump must be taken out of service and disassembled to do
this. This is in contrast to the open impeller pump described above, which
can have its leakage joint adjusted without disassembling the pump.
A closed impeller pump may have a wear ring on either the impeller, the
casing, or both (Dual wear rings, see Figure 4.41). If there is only one wear ring,
the ring material is selected to be softer than the part it is running against, so
that the wear ring will wear, rather than the impeller or casing. Dual rings, if
they are supplied in chrome stainless, are usually heat treated to maintain a
50 Brinell hardness difference between the impeller and casing rings.
It may be possible to reduce the initial cost of a pump by purchasing it
without wear rings, and then adding them at the first maintenance outage
of the pump. Normally, this is false economy because the pump casing and/
or impeller would have to be remachined to accept the rings during the outage, and this operation would likely cost more when done as part of a pump
repair than if supplied that way by the manufacturer.
Closed impeller wear ring clearances vary according to the diameter of the
impeller or the diameter of the leakage joint (impeller suction side hub diameter). Table 4.1 shows a selected comparison of recommended wear ring clearances from two resources, the Pump Handbook (Ref. [3]) and API 610 (Ref. [8]).
Another guideline for wear ring clearances calls for a diametral clearance of
0.001 to 0.0015 in/in impeller diameter. For example, a 10-in. diameter impeller would have a diametral wear ring clearance of 0.010 to 0.015 in. Wear ring
clearances should be opened up from the recommended amounts by about
0.002 to 0.005 in, depending on the impeller size, for galling materials such as
316 stainless, for operating temperatures over about 500°F, and for horizontal
multistage pumps (to account for sag of this relatively long rotating assembly).
46
Pump Characteristics and Applications
TABLE 4.1
Wear Ring Clearances for Closed Impellers
Diameter of
Impeller Hub (in)
2
6
10
15
20
25
35
50
API 610 (Ref. [8])
Pump Handbook (Ref. [3])
Minimum Diametral
Clearance (in)
Minimum Diametral Clearance
and Tolerance (in)
0.011
0.018
0.022
0.027
0.032
0.037
0.046
0.061
0.012 ± 0.002
0.016 ± 0.002
0.018 ± 0.003
0.020 ± 0.003
0.022 ± 0.003
0.022 ± 0.003
0.026 ± 0.003
0.030 ± 0.005
Generally speaking, a closed impeller pump is more efficient than an open
impeller pump of the same size and specific speed. There are exceptions
to this rule, however, and there are several reasons for this inconsistency.
The area of the leakage joint for a closed impeller is smaller than the corresponding annular leakage space on an open impeller. This would seem to
mean that closed impellers are more efficient. Many vertical turbine pumps
(Section XI), which offer both open and closed impeller designs, have a more
efficient open impeller design. One reason is the fact that vertical turbine
pumps may have the impeller setting even tighter than the standard 0.012
to 0.015 in discussed above, at least when the pump is new. Seizure of the
rotor is less likely with an open impeller leakage joint than with an axially
oriented one such as in the closed impeller described above. Another point
is that liquid flowing through an open impeller is exposed to a machined
surface on one side (the machined casing contour), while in a closed impeller,
the liquid is exposed to an as-cast surface on both sides.
An alternative wear ring design, know as serrated rings (Figure 4.2), relies
on machined grooves in the rings to provide additional pressure breakdown
as liquid leaks across the rings. Volumetric losses can be minimized without
reducing ring clearances to the point where galling is a concern.
A closed impeller generally produces lower axial thrust than an open impeller of roughly the same rating and specific speed. The difference in thrust
(which is explained in more detail in Section II.D to follow) between closed and
open impellers is generally quite significant, commonly 30% to 40%.
Closed impellers are almost always used with multistage pumps. The first
two reasons for this are the higher efficiency and lower axial thrust of closed
impellers. Even more importantly, with most multistage pumps each of the
impellers is fixed in its location on the shaft by means of a split ring and groove
arrangement that is used to transmit the axial thrust created by the impeller
to the thrust bearing. If a multistage pump had open impellers and only one
impeller were replaced, it would be impossible to reset all of the impellers to
47
Centrifugal Pump Types and Applications
PHIGH
Rotating
wear ring
PLOW
Stationary
wear ring
FIGURE 4.2
Serrated wear rings minimize leakage to reduce efficiency losses without tightening ring clearances.
a tight leakage joint clearance. For these reasons, most multistage pumps have
closed impellers. Exceptions to this rule are vertical turbine pumps (Section XI)
and some very small radially split multistage pumps (Section VII.C), both of
which may have methods of attaching the impeller to the shaft other than the
split thrust rings just described. On these types of pumps, the impeller location is not fixed with respect to the shaft. Therefore, these pumps are able to
be offered by some manufacturers in multistage open impeller configuration.
The closed impeller designs discussed thus far are best suited to relatively
clean, noncorrosive liquids. They are especially suited for high-temperature
applications. The axially oriented leakage joint allows axial movement due
to temperature expansion without affecting the rate of leakage through the
leakage joint. On the downside, the axially oriented leakage joint may clog
with stringy solids, wears rapidly with erosive solids, and is more likely to
gall when furnished in stainless steel.
Many abrasive slurry pumps use another type of closed impeller, which
is illustrated in Figure 4.32. This impeller type is designed for clogging solids and abrasive and corrosive slurries. The design uses a leakage joint on
the front side of the impeller that is radial to the shaft, more like an open
impeller. The impeller may be rubber lined, made of hard metal for abrasion
resistance, or made of corrosion-resistant metal. The impeller is often accompanied by a replaceable wear plate that is mounted in the casing on either
the front or back side of the impeller, or both. These impellers are usually
48
Pump Characteristics and Applications
adjustable for wear, in some designs without shutting down the pump. See
further discussion of slurry pumps in Section X to follow.
From a manufacturing perspective, open impellers require less machining
because they do not have the front shroud. On the other hand, machining
the front edges of the vanes is an interrupted cut, which may require slow
machine tool speeds. Open impeller vanes can be hand-dressed and more
easily underfiled (Section II.E) to improve performance, because the vanes
are not covered by the shroud on the front side.
In summary, there are many factors affecting the right choice of impeller
type for a given pump type and application. Because of their usual higher
efficiencies and lower axial thrust, closed impellers are normally preferred,
provided that the liquid is reasonably clean and free of solids and abrasives.
If there are solids and abrasives in the pumped liquid, open impellers or
closed impellers with radially oriented leakage joints may be more suitable.
Sanitary (hygienic) centrifugal pumps, such as used in a dairy or biopharmaceutical plant, are designed with open impellers. This is primarily to allow
the pump impeller to be more thoroughly hosed down (after removal of the
front, of the casing) to prevent bacteria growth.
Some open impellers are referred to by some pump manufacturers as
semi-open or semi-enclosed. A semi-open impeller has a full or partial shroud
on the back side of the impeller vanes. The open impeller shown in Figure
4.1 has such a partial shroud and thus, in the strictest sense, is a semi-open
impeller. A pure open impeller, strictly speaking, would have only vanes
joined together by a hub at the center. Practically speaking, most open impellers must have at least a partial back shroud to enable machining of the vanes
without bending or breaking them off. The larger the shroud on the back
side of the impeller, the greater the amount of thrust generated by the impeller. (Refer to the discussion on impeller thrust in Section II.D below.)
B. Single versus Double Suction
Figure 4.3 illustrates the difference between single and double suction
impellers. Single suction impellers, both open and closed, are by far the most
common type of impeller. With this impeller type, liquid enters the impeller
from one side only.
A double suction impeller is a special type of impeller that has liquid entering
the impeller from both sides. Double suction impellers are usually associated
with horizontally split case pumps, either in single stage (Section VI) or as the
first stage in multistage (Section VII) configuration. They are also supplied in
vertical configurations, either as a single stage, or as the first stage of vertical turbine pumps (Section XI). Double suction impellers are always closed impellers.
There are two primary benefits of double suction impellers compared
with single suction impellers. The first benefit is that the axial thrust of a
double suction impeller is much less than that of a single suction impeller,
and very nearly zero. (Refer to further discussion of how impellers generate
49
Discharge pressure
Discharge
pressure
Suction
pressure
Suction
pressure
Suction
pressure
Discharge
pressure
Discharge
pressure
Centrifugal Pump Types and Applications
Double suction
impeller
Single suction
impeller
FIGURE 4.3
Single and double suction impellers and thrust development. (From Karassik, I.J. et al., Pump
Handbook, 4th Ed., McGraw-Hill, Inc., New York, 2008. With permission.)
thrust in Section II.D to follow.) Use of a double suction impeller means that
thrust bearing loads are lower, or that the pump may not even need a thrust
bearing.
The second benefit of a double suction impeller is that the NPSHr of a
double suction impeller is much lower than that of a comparably sized single
suction impeller with the same type of inlet design. This is true because each
of the two inlets on the double suction impeller has only one half of the flow
going into it. (See Chapter 2, Section VI for discussion of NPSH.)
Double suction impellers are often used in conjunction with double volute
casings. The double volute casing (see Chapter 1, Section V) is there to reduce
the radial bearing loads on the pump, and is usually of most benefit in higher
flow pumps where radial bearing loads are likely to be higher. The double
suction impeller is present to reduce the NPSHr, which is also more likely to
be a problem at higher flows.
C. Suction Specific Speed
Suction specific speed, a nondimensional index used to describe the geometry
of the suction side of an impeller, was first explained in Chapter 2, Section
VII. The suction specific speed is designated Nss or S, and its formula, very
similar to the formula for specific speed (Chapter 2, Section VII), is
S =
N× Q
NPSH 3r /4
(4.1)
As with specific speed, the terms in Equation 4.1 are taken at BEP, full diameter. The value of Q in Equation 4.1 for a double suction impeller (Section II.B)
is taken as one half the total pump capacity.
50
Pump Characteristics and Applications
Typical values of S for most standard designed impellers lie in the range
of 8000 to 9000. The discussion of NPSH in Chapter 2, Section VI.C, mentions special types of single suction impellers that have enlarged inlets to
reduce the NPSHr. These large eye single suction impellers typically have a
value of S in the range of 10,000 to 13,000. Caution should be used in applying high suction specific speed impellers because the acceptable range of
operation on the pump curve may be more restricted with pumps having
higher S values. Operating outside this safe range may cause the pump to
operate unstably, have excessive recirculation, and may cause cavitation,
noise, and vibration. The reasons for this phenomenon are described in
Chapter 8, Section III.B.
Another example of a high suction specific speed impeller is an inducer.
An inducer is a special type of impeller that looks almost like a screw having only a few threads (Figure 4.4). It is usually used directly in front of a
standard impeller to reduce the value of NPSHr of the standard impeller.
The value of suction specific speed for an impeller with an inducer is typically 18,000–20,000. This means, considering the discussion of the previous
paragraph, that an inducer has an even more restrictive acceptable range of
flow than a comparable impeller without one.
Section II.B above discusses the benefits of a double suction impeller, including the lowered value of NPSHr. Because the flow into a double
suction impeller is split in two, and only half the flow goes into each inlet of
the impeller, this means that the double suction impeller is able to achieve
the improved NPSHr without a higher suction specific speed. Therefore, the
double suction impeller is not subject to the same restrictive flow range as is
a large eye single suction impeller or an inducer.
FIGURE 4.4
Inducers are used to reduce NPSHr. (Courtesy of Goulds Pumps, Inc., a subsidiary of ITT
Corporation.)
Centrifugal Pump Types and Applications
51
D. Axial Thrust and Thrust Balancing
Axial thrust is caused by pressure acting against the cross-sectional area of
an impeller. This thrust must be accommodated by the pump’s thrust bearing.
Figure 4.3 shows the thrust profile for a single suction closed impeller and a
double suction impeller. With the single suction impeller, the back side of the
impeller sees nearly full discharge pressure. The front side is subject to neardischarge pressure outside the leakage joint and near-suction pressure inside
the leakage joint. Thus, there is a net thrust toward suction, the ordinary direction of axial thrust for a single stage pump having a single suction impeller.
In an open impeller, the pressure on the front side of the impeller breaks
down from discharge to suction across the leakage joint. On the back side of
the impeller, the pressure may be nearly discharge pressure across the entire
back face, or may be lower than this closer to the hub if the impeller is fitted
with pump-out vanes (described below). This results in a net axial thrust that
is typically 30%–40% higher than that of a comparably sized closed impeller.
With the double suction impeller in Figure 4.3, because flow evenly splits
between the two impeller inlets, the thrust forces on each side of the impeller cancel out each other. This results in a complete balancing of axial thrust,
provided that the flow is evenly split between the two impeller inlets. Many
larger double suction pumps have a thrust bearing to account for the possible uneven distribution of the flow between the two impeller inlets or due
to the preference by the manufacturer for the impeller to have a positive
axial thrust in one direction instead of floating.
Because thrust in a centrifugal pump impeller is a function of the differential head as well as the area under pressure, thrust varies depending
on where the pump is operating on its curve. The farther to the left of BEP
a pump operates, the higher will be the thrust generated and the load the
thrust bearing must carry. This may be a factor limiting minimum flow, as is
discussed in Chapter 8, Section III.B.
There are a number of ways to balance or reduce the amount of axial thrust
generated by a single suction impeller. One such method, very commonly used
with enclosed impellers, is illustrated in Figure 4.5. The pump in this figure has
a second wear ring located on the back side of the impeller. Additionally, several small holes have been drilled in the impeller, permitting a leak path from
the back side of the impeller to suction. High-pressure liquid from the pump
discharge leaks across the back wear ring, reducing in pressure as it throttles
across the ring. Then the liquid, when it reaches the cavity at the back side of
the impeller, is exposed to the leak path through the drilled holes, and leaks
back to the suction side of the impeller. The net effect of these two features on
the pump is that the pressure in the cavity on the back side of the impeller is
reduced to nearly suction pressure, thus reducing the axial thrust as well as
lowering the pressure that must be carried by the seal.
This thrust balancing arrangement has a cost, naturally. First, there is additional machining and an additional component (the wear ring on the back
52
Pump Characteristics and Applications
Discharge pressure
Front
wearing
ring
Back
wearing
ring
Balancing
hole
Area A
Area A
Suction
pressure
Suction
pressure
FIGURE 4.5
Axial thrust is balanced by the addition of a back wear ring and holes drilled in the impeller.
(From Karassik, I.J. et al., Pump Handbook, 4th Ed., McGraw-Hill, Inc., New York, 2008. With
permission.)
side of the impeller) required when the pump is built. Second, the clearance
of the back wear ring must be regularly maintained if the thrust balance is to
remain effective. Otherwise, if the wear ring clearance were allowed to open
excessively, the liquid could not relieve through the drilled holes fast enough
and the pressure on the back side of the impeller would increase, raising
thrust and stuffing box pressure. Finally, this feature causes an additional
loss of efficiency because more liquid is leaked back to suction rather than
being pumped through the pump. These costs are often deemed acceptable
in return for the greatly reduced size of thrust bearing that must be used in
the pump, as well as the reduction in stuffing box or seal housing pressure.
Note that a few pump manufacturers use a more simplified version of the
above, having the holes drilled through the impeller, but not having the back
wear ring. This system is not as effective at balancing thrust as the system
described above, and causes even more efficiency losses, but is simpler and
less expensive to construct.
Another means of reducing axial thrust, commonly found on open impellers, is through pump-out vanes. These vanes are on the back side of some
open impellers, have a width of only 1/16 to 1/8 in, and usually follow the
contour of the impeller vanes. The pump-out vanes produce a pumping
action on the back side of the pump impeller as it rotates, pushing liquid out
of the area behind the impeller. This reduces the pressure on the back side of
the impeller, lowering the axial thrust. An additional (and sometimes more
important) benefit of back pump-out vanes is that, with the pressure on the
back side of the impeller reduced, the packing or mechanical seal must seal
against a lower pressure, which usually means a longer packing or seal life.
Centrifugal Pump Types and Applications
53
Still another benefit of the back pump-out vanes is to keep solids and abrasives out of the packing or seal area, a common application for slurry pumps.
(Refer to Section X.)
Note that as the open impeller wears and is adjusted axially to reestablish
higher efficiency, this reduces the effectiveness of the back pump-out vanes
(because the clearance on the back side of the impeller would have increased
from the impeller adjustment). This, in turn, results in higher stuffing box pressure and higher axial thrust. Therefore, prior to impeller adjustment, users
should consider the trade-off between higher flow and higher pump efficiency
(lower pumping costs) on the one hand, vs. higher thrust load and seal pressure
(possibly higher maintenance costs) on the other hand. Each situation is unique,
of course, and due consideration should be given to other factors such as the
conservatism in the bearing design and the pump operating point or duty cycle.
Under conditions of high suction pressure, especially with an impeller that
has been thrust balanced as described above, the thrust may reverse and can
result in a high thrust toward the stuffing box, opposite of the normal direction of thrust.
Another type of thrust balancing device, common with multistage pumps
that do not have opposed impellers (see Section VII), is a balance drum. This
device, shown on the pump in Figure 4.23, has a sleeve or drum attached
to the shaft at the high-pressure end of the pump. The drum is exposed to
full discharge pressure on the front side, and a much lower pressure on the
back side, creating a thrust counter to the normal direction of thrust. The
balance drum runs in a long throttle bushing that might be serrated or have
a labyrinth design to provide the maximum amount of pressure breakdown.
This large amount of pressure breakdown makes these devices rather highmaintenance devices, but this may be the only feasible way to balance axial
thrust in some types of multistage pumps.
E. Filing Impeller Vane Tips
Impeller vane tips can be filed on both the inlet and the outlet edges of the
vanes as a way to sometimes improve pump performance. Figure 4.6 illustrates
the location of filing operations. Impeller outlet tips are underfiled by removing a certain amount of material from the under side of the vanes at the tip, and
then filing back several inches to achieve a smooth contour. This has the effect
of increasing the normal dimension between adjacent vanes (from d to dF in
Figure 4.6), as well as changing the angle of the liquid exiting from the impeller.
This often results in a slight improvement of performance, including the possibility of an increase of several percent in flow, head, or efficiency of the pump.
The results of underfiling the vane outlets are not consistent from one type
of pump to another, as tests by manufacturers on a variety of impeller sizes
and specific speeds have demonstrated. However, once underfiling has been
shown to be effective for a particular impeller, future underfiling of impellers
made of the same material and with the same drawings will have repeatable
54
Pump Characteristics and Applications
C
A
d
d ≅ dF
Removed by
overfiling
dF
Removed by
underfiling
Overfiled
E
Underfiled
D
B
FIGURE 4.6
Underfiling of vane outlets may improve performance. Filing vane inlet edges may improve
NPSHr. (From Karassik, I.J. et al., Pump Handbook, 4th Ed., McGraw-Hill, Inc., New York, 2008.
With permission.)
results. Therefore, underfiling is sometimes considered by a manufacturer
at the time that a newly developed pump is being tested, especially if the
pump’s early test results do not come up to the manufacturer’s expectations.
In that case, if the underfiled impeller performs much better, the manufacturer may elect to add the underfile to the bill of material, so that all future
impellers made to that same drawing number are underfiled as well. Or, the
manufacturer may simply modify the pattern equipment to make the underfile a part of the casting. Sometimes, an underfile is done by a manufacturer
to achieve quoted efficiencies. The decision to do an impeller underfile is not
made lightly by the manufacturer and is only done if the improvement of
pump performance justifies the additional labor required to do the job.
Underfiling can also be done in the field if it is desired to slightly increase
the flow or head of an mpeller, or if the efficiency must be improved slightly
to keep a motor from being overloaded. Before a user attempts an underfile,
it is a good idea to consult with the manufacturer to see if the manufacturer
has attempted an underfile on this impeller in the past and knows what to
expect in the way of performance change.
Underfiling the impeller outlet vane tips must be done on all of the vanes
and on the under side only. Referring to Figure 4.6, overfiling of impeller
outlet tips has not proven effective in improving pump performance. Also,
the user must realize that the process of underfiling is destructive to the
impeller. That is, if the user is not happy with the results, the process cannot
be reversed because it involves removing metal from the impeller. Following
an underfile, the impeller must be rebalanced too.
Underfiling is normally done by hand, using a pencil grinder. It is not
practical to underfile smaller, low specific speed, closed impellers because
Centrifugal Pump Types and Applications
55
the vanes are so narrow that a pencil grinder cannot reach in to do the filing operation. The practicality of underfiling also depends on the impeller
material because the removal of metal must be done by hand. For example,
if the impeller material is stainless steel and the impeller is any size at all,
an underfiling operation would take so long to perform that it might not be
worth the effort.
Filing of impeller inlet vane tips, also shown on Figure 4.6, can sometimes
be done to achieve a reduction in the NPSHr of the pump. This improvement
can be achieved if the impeller vane inlet edges are blunt or very rough.
F. Solids Handling Impellers
Section II.A above discussed the fact that open impellers can handle liquids containing solids better than closed impellers. When solids sizes are
larger than about 1 in, or if the fluid contains stringy material or rags, special impeller designs are used to handle these larger solids or the solids are
chopped by the pump. The special impeller types include nonclog, vortex and
chopper impellers.
Nonclog impellers are designed to accommodate large solids without clogging the pump. The impeller may be of the open or enclosed type, usually has
a minimal number of vanes (often two or three), and the vanes are designed
with a width and curvature to allow solids to pass through the pump. These
impellers are usually in the mixed flow specific speed range, designed to
handle high capacities with relatively low heads. The largest impellers of this
style have capacities up to 50,000 gpm, heads up to several hundred feet, and
are capable of handling solids up to 6 in diameter.
Nonclog impellers come in a number of pump configurations, depending
on the details of the installation, many of which are shown later in this chapter. For sewage applications, they can be supplied in horizontal end suction
or in vertical dry pit installations. They can also be supplied in a vertical
wet pit column arrangement, or in a self-priming arrangement, where the
pump is located above the wet well (the latter is shown in Figures 4.14 and
4.15). Another nonclog option for sewage and industrial waste services that
is growing in popularity is the submersible configuration (Figure 4.29).
Another impeller type designed to handle large solids and stringy material is a vortex impeller, which is illustrated in Figure 4.7. Also called a
recessed impeller, this impeller type is an open impeller, but with a large
space between the front edges of the impeller vanes and the casing. The
pumped liquid is induced into the pumping chamber by the vortex created
as the impeller rotates. Thus, the liquid passes through the pump without actually coming in direct contact with the impeller. This impeller type
is quite inefficient, but may be a good alternative for handling solids or
stringy material. Like the nonclog impeller, the vortex impeller also comes
in a variety of configurations (e.g., end suction, submersible, and vertical
column type).
56
Pump Characteristics and Applications
FIGURE 4.7
Vortex impeller pump. (Courtesy of Goulds Pumps, Inc., a subsidiary of ITT Corporation.)
Some solids handling pump types have a chopper blade in front of the impeller, to chop waste before passing it through the impeller. Sewage and industrial
wastewater often contains floor wipes, rags, hair, rope, etc., which wind together
to make matted balls in the wastewater and which are prone to clogging. Figure
4.26 shows a chopper pump in a vertical column sump arrangement, but chopper pumps can be dry pit vertical or end suction types too. Chopper blades are
used in residential submersible grinder pump applications too (see Figure 4.30).
III. End Suction Pumps
A. Close-Coupled Pumps
Figures 4.8 and 4.9 illustrate an end suction close-coupled centrifugal pump,
by far the most common type of centrifugal pump. This pump type is called
an end suction pump because the liquid enters the pump from the end, with
the discharge being at a right angle from the shaft. The pump type is referred
to as close-coupled because of the fact that the impeller is directly connected
to the motor shaft, rather than being separated by a shaft coupling. With a
close-coupled pump, the pump casing (almost always a single volute casing)
is directly attached to the end face of the motor (or separated by a connecting bracket). There are no separate bearings, either radial or thrust, in the
Centrifugal Pump Types and Applications
57
FIGURE 4.8
Close-coupled end suction pump. (Courtesy of Crane Pumps & Systems, Inc., Piqua, OH.)
FIGURE 4.9
Close-coupled end suction pump. (Courtesy of Crane Pumps & Systems, Inc., Piqua, OH.)
pump. All radial and axial thrust loads must be supported by the bearings in
the motor. Because there are no separate bearings for the pump, most makers of this type of configuration usually limit their offering to about 60 HP
(although several manufacturers go up to 100 HP). Still, this covers a huge
majority of centrifugal pumps produced.
58
Pump Characteristics and Applications
The close-coupled end suction pump is popular because of its compactness, its simplicity, and the fact that it is the lowest cost configuration for a
single stage pump. A great many U.S. companies make this type of pump
in standard materials such as iron and bronze. A number of pump manufacturers offer the pump in stainless and other alloys, as well as in plastics.
Close-coupled pumps are usually fitted with closed impellers because of the
difficulty in adjusting axial clearances with this type of pump. The thrust
bearing for the motors used with this pump type is often not limited to 0.001
in axial movement, as is necessary with open impellers, but instead usually
allows more than 0.005 in movement. Close-coupled pumps are quite often
fitted with a single mechanical seal. The pump may or may not have wear
rings. (Wear rings are not included in the pump in Figure 4.9.) Upper temperature limits are usually about 225°F, and the most common service is for
ambient or relatively low temperature water.
Close-coupled pumps normally come in sizes up to 4 or 5 in discharge,
with flows up to about 2000 gpm and heads up to about 700 ft.
Note that when maximum flows and heads are given in this chapter for particular pump types, it should be understood that these maximum values are
not available concurrently. A pump achieving the highest flow listed would
normally be run at a slower speed than 3600 rpm and would have a much lower
head than the maximum head listed. The interested reader should look at several manufacturers’ catalogs to ensure that a particular rating can be made.
Suction and discharge connections for end suction close-coupled pumps
are often threaded up to about 2 or 3 in and flanged in larger sizes. The usual
60-cycle operating speeds for this pump type are 3600, 1800, and 1200 rpm.
Close-coupled pumps require a C-Face motor, different from the footmounted motors used with most other pump types. C-Face motors have the
inboard end machined to dimensions standardized by NEMA, the National
Electrical Manufacturers’ Association, so that pump manufacturers can easily adapt their pumps to them.
In addition to their relatively low cost, compactness, and simplicity, closecoupled pumps also enjoy the distinct advantage, compared with the framemounted pump design discussed in the following section, of not requiring
coupling alignment. The rabbet fits of the motor and pump volute ensure the
concentricity of these two components.
Because of the advantages cited above, close-coupled pumps are the most
popular type of pump for general light-duty services. They are also favored by
OEMs (original equipment manufacturers), companies that make a machine
or system that incorporates a pump, and by many commercial users.
Industrial users of pumps have historically not favored close-coupled
pumps, preferring instead the more common foot-mounted motor, which is
used on frame-mounted pumps (see Section III.B to follow) and on many
other types of industrial rotating machines (blowers, mixers, etc.). Reasons
for the lack of full acceptance of close-coupled pumps include the relatively
low upper limitations on horsepower, flow, head, and temperature in the
Centrifugal Pump Types and Applications
59
close-coupled configuration and the fact that very few special features are
available for this pump type. API 610 does not permit this type of configuration for hydrocarbon services. One other shortcoming of close-coupled
pumps is that to access the impeller or mechanical seal for maintenance,
the motor must be moved, which is not required on frame-mounted pumps,
provided the right type of coupling is chosen. Also, many years ago, C-Face
motors were not competitively available in the variety of enclosures available with foot-mounted motors. Today, C-Face motors are readily available in
all types of enclosures and at prices competitive with foot-mounted motors.
B. Frame-Mounted Pumps
The end suction frame-mounted pump, illustrated in Figures 4.10 and 4.11, differs from the close-coupled pump described in Section III.A above in that the
pump and motor (which is foot mounted) are separated by a shaft coupling.
The pump has its own bearing frame, with a radial and thrust bearing, and the
pump and motor are usually mounted on a common cast or fabricated bedplate.
Frame-mounted pumps use motors that are the same as those used on
other rotating equipment such as fans, blowers, mixers, etc. In a framemounted pump, shaft deflection, runout, and thrust-carrying capability are
controlled and designed by the pump manufacturer.
This type of pump, because it has a separate bearing frame, can be made
in much larger sizes than close-coupled pumps, and is therefore more commonly used in heavy-duty industrial applications (see Figures 4.40 and 4.41).
Also, the end suction configuration is simpler and has a lower first cost than
any other single stage alternative. Some manufacturers offer this pump configuration up to flows of about 5000 gpm, and several offer it in much higher
FIGURE 4.10
Frame-mounted end suction pump. (Courtesy of Grundfos.)
60
Pump Characteristics and Applications
FIGURE 4.11
Frame-mounted end suction pump. (Courtesy of Grundfos.)
flows (to about 25,000 gpm). The casing is usually offered in single volute up
to about 2000 gpm and may be double volute for higher flows.
Because of its use in many process applications that involve high pressures, high temperatures, and corrosive liquids, frame-mounted pumps are
offered in a broad range of materials and with a variety of sealing options.
Provisions are often made to cool the bearing housing and seal housing for
high-temperature applications.
Many end suction close coupled and frame-mounted pumps are equipped
with shaft sleeves (see Figures 4.40 and 4.41). The sleeve is often sealed against
the impeller with an O-ring or gasket. Its primary function is to isolate the
shaft from the pumped liquid, so that the shaft is not exposed to the potentially corrosive and erosive liquid and to protect the shaft from wear at the
seal or packing. Therefore, the shaft can be made of a less-corrosion-resistant
alloy. Also, if there is a corrosive or abrasive attack, it is the less complex (and
therefore usually less expensive) sleeve that needs to be replaced, rather than
the more complex and expensive shaft.
Bearing lubrication for frame-mounted end suction pumps is generally
accomplished either by grease or by oil. Figure 4.11 shows grease lubricated
bearings, which is the more common lubrication system employed in lighterduty commercial pumps such as those used in commercial buildings for hot
water and HVAC systems. The bearings are greased through a grease fitting
(shown in Figure 4.11).
61
Centrifugal Pump Types and Applications
Bottle and
dust cap
assembly
Air path
Oil level
Leveling bar
FIGURE 4.12
Constant level oiler maintains oil level in bearing housing. (Courtesy of Goulds Pumps, Inc., a
subsidiary of ITT Corporation.)
Oil lubrication is the more common system for industrial frame-mounted
centrifugal pumps (regardless of the configuration). The pumps in Figures
4.40 and 4.41 show oil-lubricated bearing housings, which typically include a
vented fill cap. Oil level may be shown by a sight glass in the side of the bearing housing. Or, the pump may include a constant level oiler (Figure 4.12),
which maintains a consistent level of oil in the bearing housing, since either
too much oil or too little oil can be detrimental to proper bearing lubrication.
The oiler bubble glass holds an inventory of oil. If oil leaks out of the bearing housing through the seals on either end of the housing, an air path is
exposed in the oiler, which allows oil to flow from the oiler bubble glass into
the bearing housing until the proper level is once again achieved. Not only
does the oiler bubble glass hold an inventory of oil, but it also allows a visual
determination when the bearing housing seals are leaking and need replacement. Since many modern industrial pumps have labyrinth seals to keep the
oil from leaking out of the bearing housing, rather than the more traditional
rubber lip seal, the use of the constant level oiler has declined.
IV. Inline Pumps
Inline pumps (Figure 4.13) are usually oriented vertically and may be closecoupled or frame mounted. The suction and discharge flanges are located
62
Pump Characteristics and Applications
FIGURE 4.13
Inline pump. (Courtesy of Grundfos.)
in line with each other and on opposite sides of the pump. There are three
advantages of the inline configuration compared with a comparably sized
end suction horizontal pump. First, an inline pump takes up less floor space
than a horizontal end suction pump of the same size. This may be important
if the installation must fit into a tight spot in a building or plant, on a ship or
offshore platform, or in a skid-mounted assembly. Second, the piping coming
into and out of the pump is simpler because the suction and discharge flanges
are in line with each other. This eliminates the necessity for the piping changing directions as it must do with an end suction pump and may also eliminate
elbows or other fittings. Finally, compared with a frame-mounted horizontal
pump, the close-coupled inline configuration usually requires no field alignment because the motor and pump are aligned by rabbet fits.
Offsetting these advantages are several shortcomings of this type of configuration. With the vertical orientation, the motor is supported by the pump
rather than by a baseplate, so larger sizes may require external support. This
design is inherently less stable from a structural standpoint because loads are
not transferred via the bedplate to the foundation as is the case with horizontal
frame-mounted pumps. Suction conditions (NPSHr) may not be as good as is
the case with an end suction pump because of the several changes in direction
the flow must take as it moves from the suction flange to the impeller inlet.
Centrifugal Pump Types and Applications
63
Finally, leakage from the packing or mechanical seal does not always freely
drip down to a collection cup where it can be piped to a disposal system, as is
the case with horizontally oriented pumps. Rather, it can collect at the stuffing
box mounting area. Although the leakage can be drained from here to waste, it
is more likely to cause corrosive damage to the pump or motor support.
V. Self-Priming Centrifugal Pumps
Chapter 1, Section V, explains that with a centrifugal pump, the pump’s suction
line and impeller inlet must be filled with liquid and vented of air before the
pump can start satisfactorily. With a pump located in a system with a positive
suction head, merely opening the valve at the pump suction and opening the
vent valves at the top of the pump casing floods the suction line and vents the
casing. If the pump is operating on a suction lift, however, the suction line must
be filled before the pump can be started, a procedure known as “priming” the
pump. Chapter 8, Section II.B, describes several alternative priming techniques.
A self-priming centrifugal pump (Figure 4.14) automatically performs this priming procedure. This pump type can be close-coupled or frame mounted.
The ability of the pump to prime itself is due to the unique design of the
pump casing. Figure 4.15 illustrates how the self-priming centrifugal pump
FIGURE 4.14
Self-priming centrifugal pump can operate on a suction lift without external priming. This
particular pump is a nonclog style, designed to handle liquids containing solids. (From Pentair
Pump Group, Inc. Copyright 2012. With permission.)
64
Pump Characteristics and Applications
Pumping
Priming
FIGURE 4.15
Illustration of how the self-priming centrifugal pump works. (Courtesy of Goulds Pumps, Inc.,
a subsidiary of ITT Corporation.)
works. The casing is double volute and acts just like any other double volute
casing when the pump is operating. When the pump is shut down, liquid
drains out of the pump and the suction line, but a quantity of liquid remains
in the lower volute of the casing. When it is time for the pump to be started
again, the lower volute acts as an intake for the impeller. As liquid moves from
the lower volute chamber to the impeller, it does two things. First, the liquid
keeps the seal parts wetted during the priming cycle so that the seal faces do
not run dry. Second, the liquid moves into the upper volute, where it forces air
out the discharge of the pump. This elimination of air from the pump creates a
vacuum in the pump suction, which begins to draw liquid up the suction line.
The process, known as the priming cycle, continues until the liquid moves up
the suction line and into the impeller eye. Then, priming is complete and the
pump once again acts like any other primed centrifugal pump.
The priming cycle generally takes from 3 to 10 min to complete, depending on the size of the pump and the volume of the suction line. Maximum
lift capability is determined by the maximum vacuum that the pump will
pull during the priming cycle. Most self-priming centrifugals are restricted
to suction lifts of no more than 20 to 25 ft of water. These pumps are usually
restricted to flows of about 3000 gpm, although several manufacturers offer
flows up to about 6000 gpm.
Centrifugal Pump Types and Applications
65
Given the dilemma of emptying a sump filled with dirty, corrosive, and
abrasive liquid, a self-priming centrifugal pump is one of the best solutions
and is most likely the alternative that minimizes maintenance headaches.
That is because with this configuration, the only components in contact
with the sump liquid are the suction pipe, pump casing, impeller, and
stuffing box (or seal housing and wetted seal parts). These are all components that can readily handle nasty liquids. There are no moving parts at
all located down in the sump, no motors that must be wetted, and no bearings in contact with the liquid. No external vacuum source is required for
the priming. The only drawbacks to the self-priming pump are that it may
be more expensive than an alternative such as a column sump pump (especially in exotic alloys) due to the large casing, and the limitations on maximum flow and suction lift. The benefits often outweigh the shortcomings,
and so this pump is popular as a tough service sump pump. It is also very
common as a dewatering pump in industrial and construction arenas. For
the latter, there are versions of this pump that are portable, skid-mounted,
gasoline engine-driven, and some with impellers designed to handle large
solids (see Figure 4.14).
VI. Split Case Double Suction Pumps
Figures 4.16 and 4.17 illustrate a horizontally split case double suction pump.
The casing configuration is called horizontal split case, but a more general
term is axial split case, that is, with the casing split along the axis of the pump
shaft. In fact, some manufacturers also make this design in a vertical orientation for savings of floor space, similar to the benefit of a vertical inline
configuration. The most common arrangement of the split case pump is with
a horizontal orientation.
The benefit of the axial split case configuration is that the mechanical
design of the casing is more structurally stable because the impeller is
supported by bearings on either side of the impeller. The pump is sometimes called a between-the-bearings design, as opposed to overhung designs
(Figures 4.8, 4.9, 4.10, 4.11, 4.13, and 4.14), where the impeller is supported by
bearings located on one side of the pump only. Also, with an axially split
case config­uration as shown in Figures 4.16 and 4.17, the suction and discharge flanges are directly in line with each other, and on opposite sides of
the pump, often simplifying the piping arrangement. The axially split casing design allows the pump casing cover to be unbolted, the bearing covers
to be removed, and the pump rotating element (shaft, impeller, impeller
wear rings, sleeves, bearings, and seals) plus the casing wear rings to be
completely removed as a unit for maintenance without having to unbolt
66
Pump Characteristics and Applications
FIGURE 4.16
Horizontal (axial) split case, double suction pump. (Courtesy of Grundfos.)
FIGURE 4.17
Horizontal (axial) split case, double suction pump. (Copyright Sulzer Pumps.)
Centrifugal Pump Types and Applications
67
the pump suction and discharge flange connections, which are located in
the lower half of the casing.
In Figure 4.17, the double suction impeller is clearly shown. The horizontal
split case design is the most common one for horizontal, single stage pumps
with double suction impellers. As previously discussed in Sections II.B, II.C,
and II.D in this chapter, the major benefits of the double suction impeller are
that the pump has a lower NPSHr than a comparable single suction impeller
(without resorting to a high suction specific speed design for the inlet), and
the fact that in single stage configuration, the thrust loads are eliminated
or significantly reduced with a double suction impeller. These are two very
compelling benefits. When combined with the previously mentioned benefits of the axially split casing configuration, they are the reasons for the
popularity of the axially split case double suction pump. Because problems
with NPSH and high thrust loads are often associated with larger pumps,
these pumps are most commonly used with higher flow applications. These
pumps commonly employ a double volute casing for flows higher than
about 1500 gpm to reduce radial bearing loads and to permit smaller-diameter shafts.
The applications for the horizontal split case double suction pump include
such services as plant raw water supply, cooling water supply, cooling
tower pumps, fire water pumps, treated water distribution pumps, pipe
line pumps, and bilge and ballast pumps. They are offered in flows up to
80,000 gpm and heads up to about 2000 ft. Material options usually include
all iron, bronze fitted (cast iron casing with bronze impellers, sleeves, and
wear rings), and all 316 stainless steel. Some manufacturers also offer other
higher alloy impeller, sleeve, and wear ring materials.
Because of the balanced axial thrust, many smaller double suction pumps
have no thrust bearing at all, but merely radial bearings on each side of the
impeller. Larger sizes often have a thrust bearing on one end. This is to
accommodate brief periods during start-up when thrust is developed by the
pump and also to account for the possibility that the flow is not evenly split
between the two impeller inlets.
Many double suction horizontal split case pumps have two shaft sleeves
that serve to locate the impeller at the proper point on the shaft (the sleeves
are usually threaded on at either side of the impeller), as well as to isolate
the shaft from the pumped liquid to protect it against corrosive and abrasive
attack.
Figure 4.17 shows impeller and case wear rings on the pump. The design
shown holds the case ring against rotation by a pin.
Double suction axial split case pumps must have two sets of packing,
or two mechanical seals, because the shaft penetrates the casing on both
sides. As Figure 4.17 illustrates, however, the packing or seals are subjected to suction pressure rather than discharge pressure. This is normally
lighter duty for the packing or seals, as the lower the pressure being sealed
against, the longer the service life of the packing or seals. This can present
68
Pump Characteristics and Applications
problems, however, when the pump is operating on a suction lift. In that
situation, the pressure inside the casing on the suction side of the impeller
is below atmospheric pressure. This presents an opportunity for air to leak
into the pump across the packing or seal (more likely to be a problem with
packing). Chapter 3, Section IV, deals with the detrimental consequences
of introducing air into a centrifugal pump. This possibility is eliminated
on a double suction pump operating on a suction lift by means of bypass
tubing, which circulates high-pressure liquid from the pump discharge
around to the stuffing box or seal housing area, where it creates a liquid
seal preventing air from leaking into the pump. This introduction of liquid
under pressure into the stuffing box also ensures that the packing or seal
faces are lubricated at all times when the pump is operating, which, as
discussed in Chapter 5 to follow, is essential for long life of the packing or
mechanical seal. (Note that this bypass tubing is not shown in Figures 4.16
and 4.17, although Figure 4.16 shows the tapped connection on the stuffing
boxes with a pipe plug in it.)
Note that several manufacturers make a radial split case version of the single stage double suction pump, as shown in Figure 4.18. The radially split
case pump design is primarily used for high-temperature and light hydrocarbon process applications and is limited to flows of about 10,000 gpm. The
configuration may be top suction, top discharge (as shown in Figure 4.18) or
side suction, side discharge.
FIGURE 4.18
Radial split case, double suction pump. (Copyright Sulzer Pumps.)
Centrifugal Pump Types and Applications
69
VII. Multistage Pumps
A. General
Multistage pumps generate the highest heads of any centrifugal pump at
normal operating pump speeds. (A special type of high-speed centrifugal
pump is discussed in Chapter 7, Section VII.) Multistage pumps have multiple impellers that operate in series. The flow moves through the pump from
one stage to the next, with a volute or diffuser section following each impeller, so that the head is increased as the liquid moves through the pump. In
most axially split case designs, the pumps have from two to as many as 15
stages. Some radially split case multistage pumps are available with many
more stages than that. Vertical turbine pumps are a special type of multistage pump, and they are discussed separately in Section XI to follow.
Multistage pumps achieve much higher heads than can be obtained from
even large-diameter single stage impellers. Also, compared with single stage
pumps at the same head and capacity, multistage pumps can achieve higher
efficiencies than single stage pumps. (Refer to Chapter 6, Section II, for a
more detailed explanation of why this is true.) Compared with positive displacement pump alternatives, multistage pumps may not be as efficient, but
are often lower priced and are almost always smoother operating and with
lower maintenance costs than positive displacement alternatives. (And as
Chapter 1 indicates, positive displacement pumps cannot achieve as high a
flow rate as multistage centrifugals.)
Applications for multistage pumps include boiler feed, high-pressure process applications, spraying systems, descaling, pressure boosters for highrise buildings, reverse osmosis, and pipe line. The following two sections
describe the most important types of multistage pumps.
B. Axially Split Case Pumps
Axially split case multistage pumps are usually horizontally oriented (so
the two terms “horizontally split case” and “axially split case” are used
interchangeably in this section). This pump type may have from as few as
2 stages (Figure 4.19) to as many as 15 stages (Figure 4.20 shows 8 stages but
could have more). Some manufacturers make horizontally split case multistage pumps in flows up to about 15,000 gpm, and these pumps can develop
heads up to about 7000 ft. Applications with hot water are typically limited
to about 1800 psi with this pump type, while pumps carrying ambient temperature liquids can handle considerably higher pressures.
The horizontally split case offers the same advantages as discussed for axially split casing configurations in Section VI, except that the suction and discharge flanges, while on opposite sides of the pump, are usually not directly
in line with each other.
70
Pump Characteristics and Applications
FIGURE 4.19
Horizontally split case two-stage pump. (Courtesy of Crane Pumps & Systems, Inc., Piqua, OH.)
FIGURE 4.20
Horizontal (axial) split case multistage pump. (Copyright Sulzer Pumps.)
Centrifugal Pump Types and Applications
71
Axially split multistage pumps are made with both volute and diffuser
casing designs. Volute designs with more than two stages are usually dual
volute. A two-stage pump may have single volute construction, with the
volute for one stage 180° opposed from the other. This achieves nearly the
same effect of balancing radial loads as a dual volute casing on a single stage
pump.
Thrust loads in a horizontally split case pump can be minimized by orienting half of the impellers in one direction, and the other half in the opposite
direction. The pumps shown in Figures 4.19 and 4.20 illustrate this. The flow
passes through half of the stages (or nearly half in the case of an odd number of stages), then crosses over to the opposite side of the pump and goes
through the remaining stages. The net effect, if there is an even number of
stages, and if the impellers are all trimmed to the same diameter, is that
thrust loads are balanced. Most multistage pumps still have a thrust bearing
to accommodate thrust imbalances at startup, to accommodate designs with
an odd number of impellers, or if the impellers are not all trimmed to the
same diameter.
Most horizontally split case multistage pumps have a center case bushing
located between the two impellers that are back to back in the center of the
pump. This bushing must maintain a tight running clearance because there
is a high differential pressure across the bushing, and the leakage across it
should be minimized to maintain the pump’s efficiency and preserve thrust
balance.
There are two stuffing boxes or seals on this type of pump, but ordinarily
one would see suction pressure and the other would see some intermediate
pressure. Therefore, there is usually a connecting line (shown on Figure 4.19)
to equalize the pressure that the two seals must seal against, to equalize the
maintenance interval for the two sets of seals.
Many multistage pumps are offered with optional large eye, first stage
impellers for improved NPSHr of the first stage (see Section II.C for a discussion of suction specific speed and enlarged inlets for impellers). The pump
shown in Figure 4.20 shows such a special first stage impeller. Some designs
also offer a double suction impeller for the first stage.
Material options for horizontally split case multistage pumps usually
include cast iron (for lower pressure) or steel (for higher pressure), 12%
chrome, bronze, or 316 stainless steel.
The horizontal multistage pump with opposed impellers is a difficult casing to seal with the complex geometry required for the crossover described
above. This relies heavily on the gasket between the casing halves doing its
job. Particularly in hot water services, a very slight gasket imperfection could
result in leakage across the casing parting plane from high pressure areas
to low pressure areas, a condition known as wire drawing, which can very
quickly cause serious damage to the casing. The radially split case diffuser
pump described in the next section does not have this case gasket complexity and sealing problem.
72
Pump Characteristics and Applications
FIGURE 4.21
Radially split case multistage pump (tie-rod design). (Courtesy of Grundfos.)
Centrifugal Pump Types and Applications
73
C. Radially Split Case Pumps
There are several types of radially split case multistage pumps. Figures 4.21,
4.22, and 4.23 show three types of radially split case pumps that have impellers and diffusers stacked together, with the entire assembly held together
by a tube (Figure 4.22) or tie-rods (Figures 4.21 and 4.23). These pumps,
also called ring-joint or ring-section pumps, are generally considered in the
United States to be lighter duty than the axially split case pumps described
in Section VII.B above. They are available in both horizontal and vertical
FIGURE 4.22
Radially split case multistage pump (tube design). (Copyright General Electric Company. With
permission.)
FIGURE 4.23
Radially split case multistage pump (tie-rod design). (Copyright Sulzer Pumps.)
74
Pump Characteristics and Applications
FIGURE 4.24
Barrel pump. (Copyright Sulzer Pumps.)
configurations. They are available with flows to about 3500 gpm and with
heads to about 5000 ft. Several manufacturers make large, heavily engineered versions of this pump for process applications, but the most common
applications for this configuration are high-pressure water booster systems,
reverse osmosis, and small boiler feed services.
Figure 4.24 shows a heavier-duty, radially split case pump known as a doublebarrel, barrel, or double-case pump. This pump is used for electric utility boiler
feed and other very heavy-duty process services. Pumps of this configuration
can achieve flows to 50,000 gpm, heads to 10,000 ft, and can handle temperatures up to about 700°F. These pumps for boiler feed service generally operate
at 5000 to 7000 rpm, and often require as much as 100 ft of NPSH. Accordingly,
they usually require a booster pump ahead of them to provide sufficient NPSHa.
The impellers and diffusers of the double barrel pump shown in Figure
4.24 are enclosed in a pressure-containing outer barrel. All of the impellers
may be oriented in the same direction, or there may be a crossover.
The first stage impeller of a radially split multistage pump may be a special lowNPSHr design, either large eye or double suction. The pumps in Figures 4.23 and
4.24 show a first-stage impeller with a larger eye area than the other impellers.
VIII. Vertical Column Pumps
Figures 4.25, 4.26, 4.27, and 4.28 show various types of vertical column
pumps, one of the most common pumps used in sump pump service, or in
Centrifugal Pump Types and Applications
FIGURE 4.25
Vertical column pump. (Courtesy of Crane Pumps & Systems, Inc., Piqua, OH.)
75
76
Pump Characteristics and Applications
FIGURE 4.26
Vertical column pump with chopper. (Courtesy of Vaughan Company, Inc., Montesano, WA.)
FIGURE 4.27
Nonmetallic cantilevered submerged column pump. (Courtesy of CAMAC Industries, Sparta, NJ.)
Centrifugal Pump Types and Applications
77
FIGURE 4.28
Drum pump. (Courtesy of Lutz Pumps, Norcross, GA.)
transferring or circulating liquid from a tank or sump. The pump is a single
stage design and is often installed as a duplex unit. A thrust bearing is located
at the top of the pump (Figure 4.25), or smaller units are close-coupled. The
pump discharges through a discharge column pipe, and the shaft is enclosed
in a central vertical column pipe. This pump type is actually just a single stage
end suction pump that has been oriented vertically, with the addition of the
long column pipe enclosing the shaft and the discharge pipe. The impeller is
immersed below the liquid level, so that it has adequate NPSHa and submergence. This pump style can have an open impeller design (typical for smaller
sizes of sump pumps), a closed type, or a nonclog style impeller (for sewage
and other liquids containing large solids) as shown in Figures 4.25 and 4.26.
Several manufacturers offer chopper attachments to vertical column sump
pumps (Figure 4.26).
The suction inlet on the casing is usually opened wider because there is no
suction piping, and the suction inlet is often covered with a screen to keep out
solids that are too large to pass through the pump. Some manufacturers offer an
adapter to the casing that allows the pump to be located next to a tank, rather than
immersed in the tank, and with the suction piping running from the tank to the
suction adapter. This configuration is often referred to as a dry pit configuration.
78
Pump Characteristics and Applications
As a sump pump, this pump has the advantage of being relatively inexpensive, of rather simple construction, and easy to take apart for maintenance.
Because it has been adapted from horizontal designs, and with the flows
being usually less than 3000 gpm, the casing is usually single volute. This
is one of the weaknesses of this design in its basic configuration, because
the pump has no radial bearings except perhaps a sleeve bearing just above
the impeller. Because the single volute construction causes radial loads, this
sleeve bearing in larger pumps is often observed to wear on one side because
the radial load tends to push the impeller to one side.
The lubrication of the sleeve bearing mentioned in the preceding paragraph
is the other weakness of this pump style. There are usually other sleeve bearings supporting the shaft at intervals of 3 ft. (This pump type is usually limited
to a length of about 20 ft.) There are several alternatives for lubricating these
sleeve bearings. In one design style, the pumped liquid is allowed to get into
the shaft column to lubricate the bearings. This may not be a good solution if
the pumped liquid is corrosive or contains abrasives, although it is acceptable
if the liquid is clean and noncorrosive. Also, the shaft can be hardened at the
sleeve bearing locations to minimize abrasive wear. Another approach (shown
in Figure 4.25) has tubing directed down the outside of the shaft column, so that
the bearings can be lubricated externally by water, oil, or grease. Although this
approach isolates the bearing from the pumped liquid, it adds to the amount of
water or oily waste that must be processed in a sump pump system.
Despite the shortcomings discussed above, this pump design is popular
because of its relative low cost and ease of maintenance. It also has another
distinct advantage of not requiring a seal. The shaft column pipe usually has
holes open to the sump (when used in its normal wet-pit configuration), so
that the pipe does not build up pressure, but relieves to the sump instead.
There is often a lip seal or packing at the top of the column pipe to keep
vapors from leaking out at this point.
This pump type is available in iron, bronze, and higher alloys for more
corrosive services. There are several makers of smaller versions of this pump
type in a cantilevered configuration (eliminating bearings in the pumped
liquid), with sealless pumping being one of their major selling features.
These smaller pumps (see Figure 4.27) are supplied in nonmetallic materials
such as CPVC, polypropylene, and PVDF, and are popular in the plating and
semiconductor industries (See Chapter 7, Section V for discussion of nonmetallic pumps.).
Section X in this chapter on slurry pumps discusses an alternative design
of the cantilevered configuration in larger sizes. These pumps eliminate the
bearings in the shaft column, and some have special impeller types to handle solids and abrasives.
Still another variation of vertical column pump is the drum pump, shown in
Figure 4.28. Drum pumps are designed to fit the bung of a standard 55-gal
drum, for emptying the drum of chemicals or other liquids. There are also
designs that can accommodate pumping out of totes, vats, intermediate bulk
Centrifugal Pump Types and Applications
79
containers, carboys, and open vessels up to 10 ft deep. Drum pumps are portable, and can be supplied with a wide range of AC and DC motors and
compressed air drives. The usual construction has centrifugal-style impellers, although several manufacturers also have offerings where the rotor is a
progressing cavity design for higher viscosity liquids. Typical drum pump
wetted materials of construction include polypropylene, PVDF, aluminum,
and stainless steel.
IX. Submersible Pumps
Submersible pumps (Figures 4.29 and 4.30) use a motor that is designed to
operate submerged in the pumped liquid. The motor is often encapsulated
and filled with oil, which is separated from the pumped liquid by a mechanical seal. These pumps are usually designed to pump sewage or industrial
wastewater from a pit or a tank, with the pumped liquid often containing solids. The simplest version of this design is the residential sump pump, located
in the basements of homes where the water table is such that water will accumulate in the basement. There are also home sump pumps of the column type
as described in Section VIII above, but this older design is less popular for
residential sump pumps, and the submersible configuration now dominates
FIGURE 4.29
Submersible pump. (Courtesy of Grundfos.)
80
Pump Characteristics and Applications
FIGURE 4.30
Grinder pump. (Courtesy of Grundfos.)
this application. There are much larger submersible models, of course, for
handling the higher flows and heads and the larger solids sizes required for
commercial, municipal, and industrial applications. Also, refer to Section XI
for a discussion of submersible versions of vertical turbine pumps.
Submersible pumps such as shown in Figure 4.29 are single stage, and the
impeller is usually a nonclog style, described in Section II.F.
Another feature often found on these pumps is a slide rail system and
quick connect discharge, which allow the pump to be able to be lowered into
a waste sump and coupled to the discharge piping with minimal time spent
in the sump by the mechanic.
The submersible nonclog pump for waste handling applications is a relatively recent pump configuration, replacing an earlier generation of vertical
bottom suction dry pit nonclog pumps. Submersible motors have improved
in reliability over the past 20 years, making this configuration more acceptable for most applications. This pump configuration is available normally
only in iron and bronze, although there is limited availability of stainless
steel for corrosive industrial wastes. Residential sump pumps often have the
impeller, volute, and motor housing made largely of plastic.
Another type of submersible pump, called a grinder pump (see Figure 4.30), is
used to grind and pump sewage in pressurized sewer systems. This relatively
new application allows for smaller sewer lines in new housing and commercial
developments, because the sewage is ground up and pumped under pressure,
Centrifugal Pump Types and Applications
81
rather than having to drain by gravity to a collection station or treatment plant.
Use of the grinder pump also means that sewage lines in new housing developments can follow the contour of the land, rather than being continuously
sloping to a collection station. Grinder pumps have higher heads than sump
and effluent pumps, and have cutting blades to grind up the sewage.
As the next section indicates, submersible pumps can also be used in abrasive applications, although with special materials and impeller types to tolerate the abrasive nature of the liquids pumped.
X. Slurry Pumps
Slurry pumps have many industrial applications where liquids containing
abrasive particles must be pumped. Applications for slurry pumps include
mine dewatering and transporting of mine slurries and mine tailings, minerals processing plants, fly ash and bottom ash sluicing in coal-fired power
plants, mill scale handling, sand and gravel sluicing in quarry and dredging
operations, paper mill waste processing, and clay slurry pumping.
The major concern in pumping these and other abrasive slurries is the
abrasive wear on the pump impeller, casing, and other wetted parts. Many
of these slurries are corrosive as well. The major considerations are the selection of the best pump materials for the service, and the design of the pump
components to resist wear or be isolated from the abrasive liquid.
The two most common material choices for slurry pumps are rubber-lined
pumps and hard metal pumps. Rubber-lined pumps (Figure 4.31) are normally
chosen if the abrasives are fine, and particularly if the abrasive particles are
more rounded in shape or are corrosive as well as abrasive. If the abrasive solids
FIGURE 4.31
Rubber-lined end suction slurry pump. (Courtesy of Weir Minerals Division, Madison, WI.)
82
Pump Characteristics and Applications
are larger in size or if they have a more jagged or irregular shape, hard metal
pumps (Figure 4.32) are the more likely choice. Most hard metal pumps are
made of a high-nickel iron, heat treated to hardness levels above 600 on the
Brinell scale. These metal pumps are so hard that the components cannot be
machined using regular machine tools, but rather must be machined using
grinders. For this reason, flanges are often cast with slots to eliminate the necessity to drill holes. In addition to high-nickel iron, other material choices for slurries that are more corrosive than abrasive are 316 stainless steel and CD4MCu.
Impellers, whether rubber lined or hard metal, are often supplied in one
size only because it is impractical or impossible to trim them to achieve varying hydraulics. Therefore, many of these pumps are set up to run at one of
several speeds, with belt drives being the most common device to achieve
this. On the family of curves for this type of pump, the curves are usually
shown at a single impeller diameter but operating at several possible speeds.
Slurry pumps are generally run at speeds of 1200 rpm and slower in an effort
to reduce excessive erosion in the high-velocity areas.
There are several configurations offered for slurry pump applications.
The most common are end suction pumps, as exemplified by Figures 4.31
and 4.32. Both versions typically use enclosed impellers with a radially
oriented leakage joint.
For wet pit applications, some manufacturers offer a column sump pump,
similar to the one discussed in Section VIII above, except that the pump is
cantilevered to eliminate any bearings in the pumped liquid. This pump
configuration is usually limited to a length of about 12 ft and must have an
oversized shaft to eliminate submerged bearings.
FIGURE 4.32
Hard metal slurry pump. (Courtesy of Weir Minerals Division, Madison, WI.)
83
Centrifugal Pump Types and Applications
Finally, some manufacturers offer a submersible configuration of a hard
metal pump, similar in its basic configuration to Figure 4.29 except having
the wetted parts constructed of hard metal alloys.
Slurry pumps come with a variety of special features, in addition to the
ones already discussed, to help them withstand the effects of abrasive wear.
Some of these pumps are equipped with vortex or recessed impellers, as shown in
Figure 4.7. This impeller is recessed so that it increases the liquid velocity while
remaining outside the main liquid flow path. This results in a very inefficient
pump, but this is often offset by the much greater resistance to abrasion and the
much larger solids and stringy material that can be handled with this design.
Other features commonly found on slurry pumps include pump-out vanes
on the back side of the impeller (described in Section II.D), replaceable wear
plates on both the suction and the discharge sides of the casing, extra heavy
bearings to withstand the shock load of pumping solids, and hard-faced
sleeves to resist abrasion at the packing area. Several manufacturers offer the
capability to have the impeller clearances adjusted “on the fly” (i.e., while the
pump is running), to re-establish hydraulic performance lost due to wear of
the leakage joint, but without the necessity of shutting down the pump.
The head and efficiency of a pump in a slurry service is reduced compared
with the performance of the pump in a clean water service. Figure 4.33 shows
0.2
0.4
1.0
2
4
10
1.1
1.5
2
50
40
30
15
Particle/impeller
ratio: d50/Di (–)
01
0.0 08
00
0.
0.0
0.0
00
6
10
5
6
0.018
0.004
0
0.0
02
0.0
20
00
4
0.0
00
3
2.65
3
Example
d50 = 0.5 mm
Di = 400 mm
d50/Di = 0.0012
S = 2.65
Cv = 30%
HR = 0.79
ER = 0.75
Cv
≤
30
40
50
4
5
6
20
%
0.00
02
0.000
1
Pump impeller
diameter: Di (mm)
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95
Efficiency ratio: ER (–)
1
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
Solids specific gravity: S (–)
Solids concentration in slurry
by true volume: Cv (%)
0.1
Head ratio: HR (–)
Solids median particle size: D50 (mm)
0.01 0.02 0.04
FIGURE 4.33
Head ratio and efficiency ratio performance adjustment of pumps in slurry service. (Courtesy
of Weir Minerals Division, Madison, WI.)
84
Pump Characteristics and Applications
a performance correction chart that permits the determination of the loss of
pump total head and efficiency due to the effects of the slurry. The factors
that affect slurry pump performance in this chart include the solids specific
gravity, the solids median particle size, the solids concentration in the slurry,
and the impeller diameter. The chart includes an example. Note that several
slurry pump performance correction methods exist among slurry pump suppliers, and there is a variability in calculated results. Users should consult
with the specific slurry pump manufacturer for a given application.
XI. Vertical Turbine Pumps
The vertical turbine pump (Figure 4.34) is a special class of pump, in a category
all of its own. It was first designed as a well pump to bring groundwater to
the surface for irrigation and drinking water. In addition to these two major
applications, the vertical turbine pump has a great many industrial applications such as transfer, booster, fire pump, cooling water, and makeup water
supply. The term “turbine” is somewhat misleading, but is used for the sake of
convention.
For many industrial applications, the vertical turbine pump offers quite
a few advantages over alternative choices (such as split case double suction
pumps). The way that they can be staged allows designs with optimum efficiency to be obtained for most applications, and they may be the only choice
for some applications. Other benefits include lower installed cost in many
cases, no foot valve or priming required, and more material options. There
are several offsetting shortcomings, however, which can lead to maintenance
problems if these pumps are not chosen carefully and maintained properly.
As shown in Figure 4.34, liquid enters the lower end of the vertical turbine
pump through the suction bell. The flow passes through one or more stages
that have either open or closed impellers and diffuser cases called bowls.
Because of its deep-well origin, economics dictate that the pump be made as
narrow as possible, to minimize the required well diameter. Consequently,
the diffuser bowls are located more or less in line with the impellers rather
than outside them, as is often the case with horizontal diffuser pumps where
the diameter is not critical but the shaft length affects rotor dynamics.
Vertical turbine pumps can have anywhere from one to about 25 stages,
providing a wide range of flow and head for this pump design. Bowls range
in size from a diameter of 4 to ~100 in, with flow rates greater than 200,000
gpm being offered by several manufacturers. Multistage vertical turbines
can achieve heads up to about 5000 ft.
The vertical lineshaft to which the pump impellers are attached runs up
the length of the pump column, which is as long as necessary to get the bowl
assembly down far enough in the liquid to achieve the minimum required
Centrifugal Pump Types and Applications
FIGURE 4.34
Vertical turbine pump. (Copyright Sulzer Pumps.)
85
86
Pump Characteristics and Applications
submergence and to have adequate NPSHa. No priming is required. The
ordinary industrial application has the pump taking suction from a pond,
lake, river, tank, or intake structure, and requires a pump length of less than
50 ft. However, there are many vertical turbine pumps installed in wells
worldwide with settings over 1000 ft deep.
At the top of the pump, above the column section, the discharge head supports the vertical motor, mounts the pump at grade, and turns the direction of
flow 90°. Lighter-duty vertical turbine pumps have threaded and coupled column sections and cast iron discharge heads, whereas heavier-duty industrial
pumps use flanged column sections and fabricated discharge heads. Pumps
for industrial services are offered in a variety of alloys for corrosion resistance, while the standard materials for water services are iron bowls, iron or
bronze impellers, iron or steel discharge heads, and steel column pipe.
There are several alternative configurations to the one shown in Figure
4.34. If it is taking suction from a source other than atmospheric pressure,
the pump can be mounted in a barrel with a special double-shelled discharge
head, as shown in Figure 4.35. This canned configuration can have a suction
pressure greater than atmospheric pressure (used as a booster pump in this
application), or below atmospheric pressure (for example, in hotwell condensate applications). A vertical turbine in a barrel may be a very economical
choice for a condensate pump, where NPSHa is usually minimal. Use of this
pump type requires that the barrel be only long enough to have the first stage
impeller low enough to achieve adequate NPSHa. Referring to the formula
for NPSHa in Chapter 2, Equation 2.19, Hs for a vertical turbine installation is
the distance from the surface of the suction vessel to the inlet of the first stage
impeller, and Hf = 0. Therefore, it may be only necessary to drill a cavity large
enough to accommodate the pump barrel, as an alternative to excavating and
building an entire pump room for a horizontal pump below the condenser.
Another configuration of vertical turbine pumps uses a submersible motor
mounted to the bottom of the pump. Figure 4.36 shows a vertical turbine
pump with a submersible motor. Usually, the motor is a specially designed
long and narrow submersible motor, capable of fitting down a well casing.
The major advantage of this configuration is that it eliminates the need of
any shaft and bearings above the bowl assembly. On a very deep setting, this
can save a great deal of money, as well as eliminate a major source of maintenance problems, as discussed below. Submersible versions of vertical turbine pumps are also more forgiving of crooked wells, since they do not have
the rotating shaft in the column section. They also eliminate the need for
packing or a mechanical seal. Finally, submersible versions of turbine pumps
tend to run quieter, since the motor is located at the bottom of the well. One
of the shortcomings of this design is that, if the submersible motor fails, the
entire pump must be pulled. Submersible motors are commonly used with
bowl sizes up to 8 and 10 in and are available in much larger sizes as well.
In Europe, submersible designs are used for most vertical turbine applications. In the United States, this configuration is becoming more popular as
Centrifugal Pump Types and Applications
87
FIGURE 4.35
Vertical turbine pump mounted in a barrel can be used for high-pressure suction sources, or
for pumping from vessels that are below atmospheric pressure. (Copyright Sulzer Pumps.)
88
Pump Characteristics and Applications
FIGURE 4.36
Vertical turbine pump with a submersible motor eliminates column shaft and bearings.
(Courtesy of Grundfos.)
Centrifugal Pump Types and Applications
89
the reliability and affordability of submersible motors continue to improve
in larger sizes. Submersible vertical turbines have made significant inroads
in the industrial markets. In addition to clean services, they are used in seawater (e.g., fire pumps and seawater lift services) for platform and dockside
applications. A special application is the horizontal installation of a submersible turbine for booster service.
One of the shortcomings of all types of vertical turbine pumps is the fact
that bearings to hold radial loads are bushings mounted in each bowl, located
in retainers at intervals of every 3 to 5 ft in the column assembly (except for
submersible versions) and at the stuffing box just below the discharge head
(again, except for the submersible version, which, as mentioned above, eliminates the stuffing box and seal). The diffuser casing design that vertical turbine pumps employ does minimize radial loads, but the pump still requires
these guide bearings at regular intervals. With the configuration shown in
Figures 4.34 and 4.35, the bearings in the bowls and column must be lubricated by the product being pumped. The same thing applies to the bowl
bearings in the submersible turbine shown in Figure 4.36. This means they
are likely to wear if the pumped liquid is abrasive and/or corrosive. To help
minimize this, a variety of material combinations are available for bearings,
as well as coatings for the shaft at bearing journal locations.
The bearings can also be worn if any of the shaft sections are not straight,
which should be carefully checked at the time of assembly of the pump. A
typical straightness tolerance is 0.0005 in per foot of shaft length.
If the many registered fits on the pump (at each bowl and column joint)
are not machined concentrically or are assembled with the looseness of the
fits all stacked up on the same side, this can also result in bearings wearing
as the pump rotates. To preclude this, the bowl assembly, which is usually
assembled horizontally on a bench, should be rotated 90° on the bench as
each stage is added during assembly.
As a bearing wears at a specific location in the pump, it tends to cause the
shaft to whip as it rotates, which further exacerbates the problem and causes
the bearing wear to spread up and down the pump to other bearings. This is
difficult to observe if it happens far down in the liquid, out of sight and hearing of the observer, thus increasing the likelihood of a problem going undetected. This progressive bearing wear can eventually cause wear rings or
impellers to rub and, in the worst case, can cause the bowl assembly to seize
up, break or twist the shaft, or cause other extensive damage to the pump. At
the least, it can cause excessive vibration, which can shorten seal and motor
bearing life and reduce hydraulic performance as recirculation increases.
Another potential problem with vertical turbine pumps has to do with the fact
that they are much more flexible than most close-coupled or frame-mounted
centrifugal pumps, making them quite a bit more susceptible to resonance than
other pump types. Resonance is a condition that can occur with pumps if
the operating speed is too close to the natural frequency of the equipment as
installed in the field. Resonance produces extremely high vibration levels in the
90
Pump Characteristics and Applications
equipment and should be avoided. Most pumps are so rigid that their operating
speed never approaches the natural frequency of the equipment as they come
up to full speed from a stopped position. Vertical turbine pumps, on the other
hand, are much more flexible, and they generally pass through a natural frequency on the way up to full speed. The natural frequency of the installed vertical turbine pump is a function of many variables, the most important of which
are pump length, weight of the bowl assembly weight, and natural frequency of
the electric motor, column diameter and wall thickness, discharge head dimensions, shaft diameter, foundation mounting system, and piping support system. Some of these variables, such as piping and foundation details and motor
natural frequency, must be communicated to the pump manufacturer by outside parties. There occasionally arises a field problem where the pump, when
first started up, ends up operating at a speed too close to its natural frequency.
This phenomenon, called resonance, causes the pump to vibrate excessively.
When the problem exists, it is usually related to the portion of the pump from
the foundation up, rather than the portion below the pump mounting flange.
The only possible solutions to a resonance problem are to change the natural
frequency of the pump by acting to stiffen or soften the pump (by substituting
certain components, welding stiffeners onto the pump, externally supporting
the pump, or changing the foundation support system) or to change the operating speed of the pump if that is possible (such as on a variable-speed system).
Note that the foregoing discussion of resonance and natural frequency is
separate from the issue of shaft rotating critical speed. This is a design issue
involving the selection by the manufacturer of correct bearing spacing and
shaft diameter for given operating speeds and thrust values to ensure that
the pump rotating speed is a safe margin from the rotating critical speed of
the shaft/bearing system.
Axial thrust loads generated by the vertical turbine pump must be accommodated by a thrust bearing. Designs of U.S. pump manufacturers use a
high-thrust vertical motor to handle the thrust loads. The normal thrust is
downward, and the thrust loads must be supplied to the motor manufacturer with the motor specification. Axial thrust increases at higher values of
TH, so the thrust loads should generally be specified as a continuous-duty
load at the normal pump operating point, as well as a short-term load based
on the pump running at or near shut-off.
Vertical turbine pumps mounted in barrels with high suction pressure
may be subject to upthrust at some operating points, with this most likely
to occur at runout flow. This condition should be carefully checked and
avoided if possible by adding hydraulic balancing devices or making system modifications. This condition should be avoided for several reasons, the
most important being that operating a vertical pump in upthrust means that
the shaft is in compression rather than tension. This makes it much more
susceptible to problems with alignment and imbalance, as well as changing
the shaft rotating critical speed. Also, some thrust bearings are not designed
for continuous operation in reversed thrust.
Centrifugal Pump Types and Applications
91
FIGURE 4.37
An enclosing tube isolates the column bearing from the pumped liquid, and lubricates the bearings
inside the tube with water or oil. (Courtesy of Goulds Pumps, Inc., a subsidiary of ITT Corporation.)
92
Pump Characteristics and Applications
European designs of vertical turbine pumps typically use a separate thrust
bearing assembly mounted between the motor and pump head to carry the
thrust, allowing the use of normal thrust vertical motors. In North America,
external thrust assemblies are used in only a few applications, most notably
in the automotive manufacturing industry.
If a submersible motor is undesirable or unavailable, pumps longer than
50 ft to water level must rely on something other than the pumped liquid
to lubricate the upper sleeve bearings in the pump column and discharge
head. Otherwise, when the pump is first started, it will take too long for the
pumped liquid to reach the upper bearings, and these bearings will overheat
and damage the bearing and/or shaft. A design option to consider as an
alternative to a submersible configuration for settings over 50 ft in length
uses an enclosing tube around the column shaft (Figure 4.37) that isolates
the column shaft and bearings from the pumped liquid. Lubricating liquid
(in older installations primarily oil, now more commonly water due to environmental release restrictions on oil) is allowed to enter the enclosing tube
at the top to provide lubrication for the column bearings. This enclosed lineshaft construction also protects the lineshaft bearings from abrasive wear.
XII. Axial Flow Pumps
This category of pump is discussed in Chapter 2, Section VII, in the discussion of specific speed. An axial flow pump is designed to deliver a very high
flow rate with a low head. Flows greater than 200,000 gpm are not uncommon, with heads usually less than 50 ft. These pumps normally come in one
FIGURE 4.38
Axial flow (propeller) pump. (Courtesy of Goulds Pumps, Inc., a subsidiary of ITT Corporation.)
Centrifugal Pump Types and Applications
93
of two configurations. One type is similar to the vertical turbine pump discussed in Section XI. With a single impeller and a vertically oriented shaft,
this pump type is used in an open body of water for such services as condenser cooling water for power plants.
Another configuration of axial flow pumps is shown in Figure 4.38. This
configuration, with the impeller mounted in a cast or fabricated elbow, is
used in closed systems where a very high flow rate must be achieved, with
the only head requirement being the piping friction losses in the closed loop.
An example of such an application is evaporator recirculation.
XIII. Regenerative Turbine Pumps
This style of pump was first mentioned in Chapter 1, Section IV.B, as a special
type of pump in the kinetic category, also known as a peripheral pump. Figure 4.39
shows a regenerative turbine pump, the impeller of which is clearly seen as
markedly different from a traditional centrifugal pump impeller. Instead of having the traditional backward-curved vanes, a regenerative turbine impeller has
radially oriented teeth or buckets, having an increasing depth with increasing
diameter. As the impeller rotates, it increases the liquid’s velocity. As the liquid
moves past the teeth, the expanding area from the increasing depth causes the
liquid velocity to decrease, achieving the change to pressure energy.
The regenerative turbine pump is capable of achieving quite high heads
in single stage construction at 3600 rpm (up to about 700 ft), and several
manufacturers make this pump type with more than one stage. Flows in
this pump type are limited to about 150 gpm. The head–capacity curve of
this pump type is quite steep, with shutoff head being two to three times
FIGURE 4.39
Regenerative turbine pump. (Courtesy of Crane Pumps & Systems, Inc., Piqua, OH.)
94
Pump Characteristics and Applications
higher than the head at the best efficiency point. Also, the horsepower curve
of a regenerative turbine pump behaves opposite to most centrifugal pumps,
with the BHP of a turbine pump increasing as flow decreases. Because of
the steep head–capacity curve and the shape of the horsepower curve, it is
usually recommended that regenerative turbine pumps be protected against
overpressurization by means of a pressure relief valve, either incorporated
into the pump or external to the pump.
The most common application for the regenerative turbine pump is small
commercial boilers. For applications where this pump type can achieve the
hydraulics, it is often the lowest cost and most compact alternative, although
not as efficient as a multistage pump. Another characteristic that sets the
regenerative turbine apart from other centrifugal pumps is that the pump
can handle up to 20% vapor or noncondensable gases in the pumped liquid
(some manufacturers claim even higher upper limits of vapor).
The greatest drawback of this pump type is that very tight running clearances must be maintained within the pump to prevent the pump hydraulic performance from rapidly deteriorating. Consequently, the pumped liquid must
be very clean and the pump will tolerate no abrasives. Most regenerative turbines have no capability to make adjustments to account for wear in the pump,
although one manufacturer offers an adjustable impeller. The tighter running
clearances make this pump run a bit noisier than standard centrifugal pumps.
Iron casings and bronze impellers are the most common material options
for this pump type. Several manufacturers offer regenerative turbines in all
bronze or all 316 stainless steel.
XIV. Pump Specifications and Standards
A. General
One of the most commonly asked questions by engineers and others who
specify or purchase pumps is: How detailed should a pump specification be,
and what is the minimum information it should contain? As with most of the
questions raised in this book, there is no absolute answer, and any answer
that is given must start with, “It depends.” As an example, for an application
requiring a relatively low flow and head (such as 100 gpm at 50 ft), where the
liquid pumped is clean water at ambient temperature, and the pump configuration will be the manufacturer’s most standard close-coupled offering,
this information, along with the motor voltage and enclosure type, is pretty
much all that is required. At the other end of the spectrum, many larger and
more complex pumps for municipalities, other government entities, or electric utilities must be specified in detail to ensure that the buying agency will
not be forced to buy equipment that is not satisfactory for the service merely
because it is the lowest bid submitted.
Centrifugal Pump Types and Applications
95
With regard to the operating conditions to which the pump is exposed, it is
safe to say that it is nearly impossible to supply too much information to the
pump manufacturer. The following information regarding operating conditions should be provided in the specification in as much detail as possible.
1. Liquid Properties
The important liquid properties to specify include the liquid type, operating temperature, specific gravity, pH, viscosity, vapor pressure, amount and
type of suspended and dissolved solids, and amount and type of abrasives
or other solids. Normal conditions and the expected range of these properties should be specified.
2. Hydraulic Conditions
Design parameters that should be specified include design capacity, total
head, NPSHa (at maximum continuous flow, and at a specified reference
point such as floor elevation), suction lift or suction head (and how these will
vary), and maximum suction pressure.
In addition to design conditions, the specification should indicate the minimum and maximum flow the pump will be required to deliver, and the
duty cycle for the different flows (i.e., other expected pump flow rates and
approximate percentage of operating time at each flow rate). Keep in mind,
however, that many manufacturers only guarantee the hydraulic conditions
at a single operating point on the pump curve unless other points are specified and specifically agreed upon by the supplier. In some cases, asking other
operating points to be guaranteed may increase the cost of the equipment, or
restrict the number of potential suppliers.
For severe services or critical installations, a system head curve (see Chapter
2, Section IX) should be developed to determine the pump duty cycle. Or,
software such as described in Chapter 3, Section III, can be used to develop
the pump operating duty cycle.
3. Installation Details
Important installation details that should be specified include the location of
the installation, elevation above sea level, unusual ambient conditions (such
as extremely high humidity or salt-laden atmosphere), type of suction vessel
or sump, and space or weight limitations.
When it comes to design details, the buyer (or specifying engineer) should,
to the extent possible, allow the bidders to offer their recommendations for
the best equipment to meet the specified service conditions. Of course, there
will be circumstances where the buyer or specifying engineer may want to
detail more specific design requirements. One reason might be if a specific
design standard is to be invoked, such as ANSI or API, which are described in
96
Pump Characteristics and Applications
the following sections, or FDA requirements for sanitary pump services. Or,
the buyer may wish to specify a particular pump configuration, materials of
construction, or design details, based on the buyer’s experience with pumps
operating in similar services. Wear rings may be considered optional by the
manufacturer on certain pumps, and so they should be specified if required.
If the purchaser lacks experience with pumps in similar services, it may be
best to allow the manufacturers to offer their recommendations for the specified service, leaving the door open to explore alternatives if the evaluation
process produces them. Where the manufacturers can offer their standard
construction, the pump’s cost is usually less than that for uniquely built equipment. The bidding manufacturers should be asked to describe their offerings,
as to configuration type, materials of construction, and design details, so that a
proper evaluation can be carried out and alternatives considered if necessary.
The type of driver for the pump should be specified, with details of construction covered if the buyer is purchasing the driver separately.
B. ANSI
The ANSI Standard B73.1 (Ref. [9], also known as ASME B73.1) for chemical
process pumps was developed more than 40 years ago. Figure 4.40 shows a
pump built to the ANSI B73.1 standard.
The standard was written to ensure that chemical process pumps built by
any manufacturer that met the standard would use the same criteria for such
design details as required casing thickness, allowable shaft deflection, bearing life, etc. Another very important goal of the standard was to standardize
the configuration and envelope dimensions of the pump for common sizes.
Thus, the standard configuration for the ANSI pump is end suction, frame
mounted, with the discharge of the pump in line with the centerline of the
casing. No matter which manufacturer is making the pump, the key envelope
dimensions of each company’s size 1 × 1 1/2 − 6 pump, for instance, are identical. The key envelope dimensions here are the shaft centerline to the discharge
flange face; the suction flange face to the coupling end of the shaft; flange
diameters; flange bolt circle; size, number, and location of flange bolts; and
shaft and keyway dimensions. This standardization of dimensions allows an
engineering consultant or user company that is designing a plant containing
hundreds of ANSI pumps to complete all details of piping and foundation
design without necessarily having chosen a supplier of the pumps.
It also allows a user to switch suppliers when replacing a pump, without
making piping, base plate, or coupling changes.
Although it is known as a chemical process pump, the ANSI pump is one
of the most versatile and widely used pumps in a variety of industrial applications. The pump can achieve flows up to 5000 gpm, with heads as high
as 750 ft at lower flow rates. Because the pump is often used with corrosive chemicals, it is readily available in an extremely broad range of material
options, probably the widest array of material options of any pump type.
Centrifugal Pump Types and Applications
97
FIGURE 4.40
ANSI pump. (Courtesy of Goulds Pumps, Inc., a subsidiary of ITT Corporation.)
Standard construction material is ductile iron, but the ANSI pump is available in 316 stainless steel, bronze, CD4MCu, hastelloy, titanium, and several
other exotic metals. A number of nonmetallic material options are also available, including FRP and Teflon® PTFE-lined versions.
While most companies that make ANSI pumps offer as many as 25 sizes of
impellers and casings for their complete line, the makers of this pump type
have modularized the design so that there are no more than six or eight stuffing box/seal housing sizes and no more than two or three bearing frames to
cover the entire ANSI line. This modularization minimizes the amount of
spare parts that must be maintained by the user in a chemical plant, which
may have hundreds or even thousands of ANSI pumps.
The ANSI pump is most commonly supplied with an open impeller, good
for both clear liquids, and liquids containing solids, adjustable for wear as
described in Section II.A above. The pump has a sleeve to isolate the shaft,
and can be equipped with a variety of mechanical seal types or in sealless
versions. Bearings are normally oil lubricated, and the bearing housing can
be externally cooled to allow pumping liquids up to 500°F. The pump is normally back pull-out design, which means that the impeller, sleeve, seal, and
98
Pump Characteristics and Applications
bearings can be removed from the back of the pump for maintenance without disturbing the suction and discharge piping connections.
C. API
The API Standard 610 (Ref. [8]) for refinery pumps was written to develop a
stringent, detailed design standard for process pumps in refinery services
and other petrochemical applications. The need for this standard stems from
the fact that the liquids being handled in refineries and petrochemical plants
are often high pressure, high temperature, and usually flammable or volatile,
so safety is a very important consideration. Furthermore, the cost of lost production time in a refinery can be considerable, as can the cost of pump maintenance. Therefore, every effort is made in the API 610 Standard to maintain
pressure integrity and reliability of the pump and to reduce maintenance
expense. This attention to design detail extends to such diverse subjects as
casing thickness criteria, nozzle loading criteria, allowable shaft deflection
and runout, bearing design, and mechanical seals.
The API 610 Standard is not written around a single configuration as is
the ANSI standard for the end suction configuration. Instead, the API 610
Standard covers a variety of pump types, including end suction, axially
and radially split case multistage, vertical turbine, and others. Because they
encompass such a variety of pump types, API 610 pumps are available over
a very wide range of capacities and heads.
The 610 Standard references a number of other API specifications, such as
API 671 for couplings and API 677 for gear units, as well as standards from
the American Society for Testing and Materials (ASTM) and other organizations. Also, there are several specific API standards covering special pump
types (e.g., API 685 for sealless pumps).
Figure 4.41 shows a horizontal end suction single stage pump designed to
API 610. Note that the pump casing is centerline supported, rather than being
supported by pump feet as is the ANSI pump discussed in Section XIV.B.
Centerline support is a requirement of API 610 for horizontal pumps with
a pumped liquid temperature of 350°F and higher. The centerline support
requires pedestals on the bedplate upon which support “wings” on the casing of the pump are mounted. This minimizes the amount of movement of
the pump centerline due to thermal expansion when pumping hot liquids, so
it is especially suitable for high-temperature applications. It also allows the
pump to operate with a relatively large temperature swing without having
to realign the pump because of thermal growth.
Pump impellers are specified in API 610 as closed impellers, to be fitted
with wear rings. The wear rings must be locked against rotation by a more
positive locking arrangement than a simple interference or press fit. Some
manufacturers use set screws or pins to secure the wear rings, while a few
companies weld the rings in place. API 610 has its own recommendations for
wear ring clearances (see Table 4.1).
Centrifugal Pump Types and Applications
99
FIGURE 4.41
API 610 covers a number of pump configurations, including the end suction single-stage
frame-mounted design shown here. (Copyright Sulzer Pumps.)
Material options for API 610 do not include as many exotic alloys for corrosion resistance as found in the ANSI standard, because most hydrocarbon
products are not corrosive. Material options include cast iron, bronze, carbon
steel, 12% chrome, and 316 stainless steel.
The API 610 Standard is considered one of the most thorough and exacting of pump specifications. Because of this, it is sometimes used as a design
specification for other heavy-duty process pumps outside the refinery and
petrochemical industries.
D. ISO
The ISO standards are making significant inroads in the United States, particularly among engineering contractors and pump manufacturers involved
100
Pump Characteristics and Applications
in major projects outside the United States. The ISO standards are broader
in scope than the ANSI or API standards, and they include such subjects
as certification of manufacturers’ quality assurance programs in addition to
equipment design standards.
ISO Technical Committee 115 (TC115) is responsible for international pump
standards. The secretariat for the U.S. Technical Advisory Group to this
committee is the Hydraulic Institute. ISO TC115 has two primary working
subcommittees. Subcommittee 1 (SC-1) covers dimensions and technical specifications, and the secretariat is the British Standards Institute. Subcommittee
2 (SC-2) covers methods of testing pumps, and the secretariat is UDMA, the
German Standards Organization.
The ISO standards reference metric units exclusively, and are based on
50-cycle electric current. Metric dimensions are generally the preferred
dimensions outside the United States. However, most developing countries
have a great deal of installed equipment such as ANSI pumps built with inch
dimensions (USCS units).
The global trend is moving toward metric dimensions, and most U.S. manufacturers with global marketing organizations supply pumps with both
metric and inch dimensions. An indication of the growing importance of
ISO in some U.S. industries is the fact that API 610 uses ISO metric standards
for mechanical seals.
XV. Couplings
The primary objectives of the shaft coupling in a pump are to:
• Connect the pump and driver shafts.
• Transmit the power (torque at a given speed) between the separate
rotating shafts of the pump and driver. (This assumes the pump is
not close-coupled.) This causes both shafts to rotate in unison and at
identical speeds.
• Compensate for minor amounts of misalignment and movement of
the shaft due to vibration or thermal effects.
For vertical pumps, because the alignment of the components is assured
through registered fits, the coupling can be a rigid type that does not tolerate misalignment, as shown in Figure 4.42. For most horizontal pumps, it
is nearly impossible to maintain perfect concentricity of centerlines of the
two shafts. Flexible couplings are used to transmit the torque, while allowing
for some amount of angular and parallel misalignment. (Refer to Chapter 7,
Section IX, for a discussion of these types of shaft misalignment.)
Centrifugal Pump Types and Applications
101
FIGURE 4.42
Rigid adjustable coupling. (Courtesy of Lovejoy, Inc., Downers Grove, IL.)
In addition to transmitting the torque and allowing for angular and parallel misalignment, flexible couplings also accept torsional shock and dampen
torsional vibration, minimize lateral loads on bearings from misalignment,
and allow for axial movement of the shafts even under misaligned conditions, without transferring thrust loads from one machine element to
another.
One of the confusing things about couplings is that there are so many
alternatives from which to choose. Most of the coupling types discussed in
this section could be applied successfully with pumps below 200 HP in size.
The selection of the optimum coupling type for a given application must
take into consideration a lot of factors, including the cost and importance of
the pump and the coupling, the pump’s duty cycle, torque to be transmitted, shaft size, permissible misalignment, temperature, and required maintenance. This section should help bring the reader a better understanding of
the various flexible coupling types to help pick the most suitable one for a
given application.
Flexible couplings are broadly divided into two major categories: metallic and elastomeric couplings. Metallic couplings have mechanical elements
such as gears (Figure 4.43), grid springs (Figure 4.44) or chains, or metallic
elements such as discs (Figure 4.45) or diaphragms. The elements allow for
some movement and misalignment of the two connected shafts. One significant feature of the mechanical element type is that many types require
lubrication.
Elastomeric couplings are further divided into two subcategories, those
that operate using compression (e.g., Lovejoy jaw type, Figure 4.46) and
102
Pump Characteristics and Applications
FIGURE 4.43
Gear coupling. (Courtesy of Rexnord Industries, Milwaukee, WI.)
FIGURE 4.44
Grid spring coupling. (Courtesy of Rexnord Industries, Milwaukee, WI.)
FIGURE 4.45
Disc coupling. (Courtesy of Lovejoy, Inc., Downers Grove, IL.)
Centrifugal Pump Types and Applications
103
FIGURE 4.46
Jaw type coupling. (Courtesy of Lovejoy, Inc., Downers Grove, IL.)
FIGURE 4.47
Bonded tire–urethane type coupling. (Courtesy of Rexnord Industries, Milwaukee, WI.)
those that operate using shear (e.g., Falk bonded tire—urethane type,
Figure 4.47). There are also some that use some combination of shear and
compression. In general, the elastomeric couplings are lower in cost and
have lower upper limits of torque and temperature than metallic couplings.
Elastomeric couplings do not require lubrication.
To help make the optimal pump coupling selection, refer to Table 4.2
(Flexible Coupling Functional Capabilities Chart) and Table 4.3 (Fexible
Coupling Evaluation Factors Chart), which have been adapted from charts
produced by a major coupling manufacturer. These charts provide the specification limits of the major coupling types, and rate them in the context of the
most important application criteria.
45.000
16.313
15.500
19.670
11.000
8.000
6.700
450,000
175,000
177,000
54,000,000
2,700,000
4,000,000
6,000,000
7.000
5.688
5.500
Maximum Bore
Diameter (in)
170,000
88,500
72,480
Source: Courtesy of Lovejoy, Inc., Downers Grove, IL.
a European standard for Jaw type couplings.
b For example, Woods Sureflex, Lovejoy S-Flex.
c For example, Dodge Paraflex, Rexnord Torus.
d For example, Rexnord Omega.
e Used for engine driven vertical turbines.
f High-speed applications.
g Manufacturers’ catalogs do not give ratings.
Metallic Types
Gear
Grid spring
Disc
Diaphragmf
Elastomeric Types
Jaw
Curved jawa
Shear type
donut—sleeveb
Corded tirec
Bonded tire—urethaned
Rubber in shear—flywheele
Flexible Coupling Type
Maximum Continuous
Torque (in-lb)
Flexible Coupling Functional Capabilities Chart
TABLE 4.2
1.50
0.25
0.50
0.50
4.00
4.00
0.50
1.00
0.90–1.30
1.00
Maximum Angular
Misalignment (degrees)
g
0.005–0.320
0.012–0.022
0.009
0.125
0.188
0.020
0.015
0.008–0.086
0.010–0.062
Maximum
Parallel Offset
(in)
0.125–1.00
0.125–0.250
0.009–0.400
0.100–0.875
0.250
0.125
0.080–0.866
0.028–0.310
0.023–0.181
0.125
Maximum Axial
Freedom (in)
104
Pump Characteristics and Applications
L
H
L
L
Shear type donut
– sleeveb
Corded tirec
Bonded tire
– urethaned
Rubber in shear
– flywheele
L
L–M
M
M–H
H
H
H
L
L
L
L
M
M
M–H
H
M
H
L
L
L
L
M
M
None
None
M
M–H
M
L
L
L–M
L–M
L–M
H
H
M
M
M–H
L–M
L–M
L
L
L
L–M
L–M
M
H
L
L
L
L
H
H
E
E
G
E
F
F
F
F
G
F
F
F
F
F–G
E
E
E
G
G
G
E
E
G
G
F–G
F
F
F
F
F
F
F
F
F
E
G
G
E
E
E
None
None
F–G
None
E
G
E
E
G
G
N
N
Y
Y
N
N
N
N
N
N
N
N
N
Y
N
N
N
N
Y
Y
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
Reactionary
High
Replace
Torsional
Relative Loads due to
Speed
Capacity for Inherent Ease of Dampening Lubrication Fail Elements in
Stiffness Backlash
Cost
Misalignment Capacity Misalignment Balance Assembly Capacity
Required Safe
Field
Source: Courtesy of Lovejoy, Inc., Downers Grove, IL.
Note: H = high, M = medium, L = low, E = excellent, G = good, F = fair, Y = yes, N = no.
a
European standard for Jaw type couplings.
b
For example, Woods Sureflex, Lovejoy S-Flex.
c
For example, Dodge Paraflex, Rexnord Torus.
d
For example, Rexnord Omega.
e
Used for engine-driven vertical turbines.
f
High-speed applications.
Diaphragmf
Disc
Grid spring
Gear
M–H
M
Metallic Types
M
Curved jawa
Axial
Torque
Forces
Capacity
Generated to O.D.
Jaw
Elastomeric Types
Flexible
Coupling Type
Flexible Coupling Evaluation Factors Chart
TABLE 4.3
5
4–8
3–5
3–5
3–5
2–3
3–5
2–3
3–5
3–5
Estimated
Service
Life (Year)
Centrifugal Pump Types and Applications
105
106
Pump Characteristics and Applications
XVI. Electric Motors
The electric motor (Figure 4.48) is one of, if not the most common machine
used in industrial and commercial settings. Many books have been written
on the subject of electric motors, but the following section should provide
an overview of electric motors for readers who are interested to learn more
about the driver type used for most pumps.
The following materials are excerpted from the Cowern Papers, written by
and with the permission of Edward Cowern, P.E., North Haven, CT.
A. Glossary of Frequently Occurring Motor Terms
Below is a glossary of some of the most frequently occurring terms related
to electric motors.
1. Amps
Full-load amps: The amount of current the motor can be expected to
draw under full load (torque) conditions is called full-load amps. It
is also known as nameplate amps.
Locked rotor amps: Also known as starting inrush or inrush current, this
is the amount of current the motor can be expected to draw under
starting conditions when full voltage is applied.
Service factor amps: This is the amount of current the motor will draw
when it is subjected to a percentage of overload equal to the service
factor on the nameplate of the motor. For example, many motors will
have a service factor of 1.15, meaning that the motor can handle a
FIGURE 4.48
Three-phase electric motor. (Courtesy of Baldor Electric Company, Fort Smith, AR.)
107
Centrifugal Pump Types and Applications
15% overload. The service factor amperage is the amount of current
that the motor will draw under the service factor load condition.
2. Code Letter
The code letter is an indication of the amount of inrush or locked rotor current that is required by a motor when it is started.
3. Design Letter
The design letter is an indication of the shape of the torque speed curve,
as defined by the National Electrical Manufacturers Association (NEMA).
Figure 4.49 shows the typical shape of the most commonly used three-phase
design letters, Design A, B, C, and D. Design B is the standard industrial duty
motor that has reasonable starting torque with moderate starting current
and good overall performance for most industrial applications. The other
designs are only used on fairly specialized applications. Design A motors
are not commonly specified but specialized motors used on injection molding applications have characteristics similar to Design A. The most important characteristic of Design A is the high pullout torque. Design C is a high
starting torque motor that is usually confined to hard-to-start loads, such as
conveyors that are going to operate under difficult conditions. Design D is
a so-called high slip motor and is normally limited to applications such as
cranes, hoists, and low-speed punch presses where high starting torque with
low starting current is desirable. Generally, the efficiency of Design D motors
280
D
Percent of full load torque
240
A
C
200
160
B
120
100
80
Full load torque
40
0
0
20
40
60
80
100
Percent of no load speed
FIGURE 4.49
Torque speed curves for three-phase electric motors. (Courtesy of Edward Cowern, P.E., North
Haven, CT.)
108
Pump Characteristics and Applications
at full load is rather poor and thus they are normally used on those applications where the torque characteristics are of primary importance.
4. Efficiency
Efficiency is the percentage of the input power that is actually converted to
work output from the motor shaft. Efficiency is stamped on the nameplate of
most domestically produced electric motors (see Section XVI.I for more details).
5. Frame Size
Motors come in various frame sizes to match the requirements of the application. In general, the frame size gets larger with increasing horsepower or
with decreasing speed. To promote standardization in the motor industry,
NEMA prescribes standard frame sizes for certain dimensions of standard
motors. For example, a motor with a frame size of 56 will always have a shaft
height above the base of 3 1/2 inches (see Section XVI.E for more details).
6. Frequency
This is the frequency for which the motor is designed. The most commonly
occurring frequency in the United States is 60 cycles, but on an international
basis, other frequencies such as 50 cycles can be found.
7. Full-Load Speed
An indication of the approximate speed that the motor will run when it is
putting out full rated output torque or horsepower is called full-load speed.
8. High Inertial Load
These are loads that have a relatively high flywheel effect. Large fans, blowers, punch presses, centrifuges, commercial washing machines, and other
types of similar loads can be classified as high inertial loads.
9. Insulation Class
The insulation class is a measure of the resistance of the insulating components of a motor to degradation from heat. Four major classifications of
insulation are used in motors. They are, in order of increasing thermal capabilities, A, B, F, and H (see Section XVI.D for more details).
10. Load Types
Constant horsepower: The term “constant horsepower” is used in certain
types of loads where the torque requirement is reduced as the speed
109
Centrifugal Pump Types and Applications
is increased and vice versa. The constant horsepower load is usually associated with metal removal applications such as drill presses,
lathes, milling machines, and other similar types of applications.
Constant torque: Constant torque is a term used to define a load characteristic where the amount of torque required to drive the machine is constant
regardless of the speed at which it is driven. For example, the torque
requirement of positive displacement pumps and blowers is constant.
Variable torque: Variable torque is found in loads having characteristics
requiring low torque at low speeds and increasing values of torque
as the speed is increased. Typical examples of variable torque loads
are centrifugal fans and centrifugal pumps.
11. Phase
Phase is the indication of the type of power supply for which the motor is
designed. Two major categories exist: single-phase and three-phase. There
are some very spotty areas where two-phase power is available, but this is
very insignificant.
12. Poles
This is the number of magnetic poles that appear within the motor when power
is applied. Poles always come in sets of two (a north and a south). Thus, the
number of poles within a motor is always an even number such as 2, 4, 6, 8, 10,
etc. In an alternating current (AC) motor, the number of poles works in conjunction with the frequency to determine the synchronous speed of the motor. At 50
and 60 cycles, the common arrangements are as given in Table 4.4.
13. Power Factor
Percent power factor is a measure of a particular motor’s requirements for
magnetizing amperage (see Section XVI.I for more details).
TABLE 4.4
AC Motor Synchronous Speeds
Synchronous Speed
No. Poles
2
4
6
8
10
60 Cycles
50 Cycles
3600
1800
1200
900
720
3000
1500
1000
750
600
110
Pump Characteristics and Applications
14. Service Factor
The service factor is a multiplier that indicates the amount of overload a
motor can be expected to handle. For example, a motor with a 1.0 service factor cannot be expected to handle more than its nameplate horsepower on a
continuous basis. Similarly, a motor with a 1.15 service factor can be expected
to safely handle intermittent loads amounting to 15% beyond its nameplate
horsepower (see Section XVI.C for more details).
15. Slip
Slip is used in two forms. One is the slip rpm, which is the difference between
the synchronous speed and the full-load speed. When this slip rpm is
expressed as a percentage of the synchronous speed, then it is called percent
slip or just “slip.” Most standard motors run with a full load slip of 2% to 5%.
16. Synchronous Speed
This is the speed at which the magnetic field within the motor is rotating.
It is also approximately the speed that the motor will run at under no load
conditions. For example, a 4-pole motor running on 60 cycles would have
a magnetic field speed of 1800 rpm. The no load speed of that motor shaft
would be very close to 1800, probably 1798 or 1799 rpm. The full-load speed
of the same motor might be 1750 rpm. The difference between the synchronous speed and the full-load speed is called the slip rpm of the motor.
17. Temperature
Ambient temperature: Ambient temperature is the maximum safe room
temperature surrounding the motor if it is going to be operated continuously at full load. In most cases, the standardized ambient temperature rating is 40°C (104°F). This is a very warm room. Certain
types of applications, such as on board ships and boiler rooms, may
require motors with a higher ambient temperature capability such
as 50°C or 60°C.
Temperature rise: Temperature rise is the amount of temperature change
that can be expected within the winding of the motor from nonoperating (cool condition) to its temperature at full load continuous operating condition. Temperature rise is normally expressed in degrees
centigrade.
18. Time Rating
Most motors are rated for continuous duty, which means that they can operate
at full load torque continuously without overheating. Motors used on certain
Centrifugal Pump Types and Applications
111
types of applications, such as waste disposal, valve actuators, hoists, and other
types of intermittent loads, will frequently be rated for short-term duty such as
5 min, 15 min, 30 min, or 1 h. Just like a human being, a motor can be asked to
handle very strenuous work as long as it is not required on a continuous basis.
19. Voltage
This refers to the voltage rating for which the motor is designed.
B. Motor Enclosures
The most reliable piece of electrical equipment in service today is a transformer. The second most reliable is the three-phase induction motor. Properly
applied and maintained, three-phase motors will last many years. One key
element of motor longevity is proper cooling. Motors are generally classified
by the method used to dissipate the internal heat.
Several standard motor enclosures are available to handle the range of
applications from “clean and dry” such as indoor air handlers, to the “wet or
worse” as found on roofs and wet cooling towers. The most common enclosure types are summarized below.
1. Open Drip-Proof
Open drip-proof (ODP) motors are good for clean and dry environments. As
the name implies, drip-proof motors can handle some dripping water provided it falls from overhead or no more than 15° off vertical. These motors
usually have ventilating openings that face down. The end housings can
frequently be rotated to maintain “drip-proof” integrity when the motor is
mounted in a different orientation. These motors are cooled by a continuous
flow of the surrounding air through the internal parts of the motor.
2. Totally Enclosed Fan Cooled
Totally enclosed fan-cooled (TEFC) motors are cooled by an external fan
mounted on the end opposite the motor output shaft. The fan blows ambient air across the outside surface of the motor to carry heat away. Air does
not move through the inside of the motor, so TEFC motors are suited for
dirty, dusty, and outdoor applications. There are many special types of TEFC
motors, including corrosion protected, chemical duty, and washdown styles.
These motors have special features to handle difficult environments. TEFC
motors generally have “weep holes” at their lowest points to prevent condensation from puddling inside the motor. As in open drip-proof motors, if the
TEFC motor is mounted in a position other than horizontal, the end housings
can generally be repositioned to keep the weep holes at the lowest point.
112
Pump Characteristics and Applications
3. Totally Enclosed Air Over
Totally enclosed air over (TEAO) motors are applied on machines such as
vane axial fans where the air moved by a direct connected fan passes over the
motor and cools it. TEAO motors frequently have dual HP ratings, depending on the speed and temperature of the cooling air. Typical ratings for a
motor might be 10 HP with 750 ft/min (fpm) of 104°F air, 10 HP with 400 fpm
of 70°F air, or 12.5 HP with 3000 fpm of 70°F air. TEAO motors are usually
confined to original equipment manufacturer (OEM) applications because
the air temperature and flows need to be predetermined.
4. Totally Enclosed Nonventilated
Totally enclosed nonventilated (TENV) motors are generally confined
to small sizes (usually under 5 HP) where the motor surface area is large
enough to radiate and convect the heat to the outside air without an external
fan or air flow. They have been popular in textile applications because lint
cannot obstruct cooling.
5. Hazardous Location
Hazardous Location motors are a special form of totally enclosed motor.
They fall into different categories, depending upon the application and environment, as defined in Article 500 of the National Electrical Code. The two
most common hazardous location motors are Class I, Explosion Proof, and
Class II, Dust Ignition Resistant. The term “explosion proof” is commonly
but erroneously used to refer to all categories of hazardous location motors.
Explosion proof applies only to Class I environments, which are those that
involve potentially explosive liquids, vapors, and gases. Class II is termed
Dust Ignition Resistant. These motors are used in environments that contain
combustible dusts such as coal, grain, flour, etc.
C. Service Factor
Some motors carry a service factor other than 1.0. This means the motor can
handle loads above the rated HP. A motor with a 1.15 service factor can handle a 15% overload, so a 10 HP motor with a 1.15 service factor can handle 11.5
HP of load. Standard open drip-proof motors have a 1.15 service factor, and
fractional HP and subfractional ODP motors can have substantially higher
service factors, in the range of 1.5 or even higher. Standard TEFC motors
have a 1.0 service factor, but most major motor manufacturers now provide
TEFC motors with a 1.15 service factor.
The question often arises whether to use service factor in motor load calculations. In general, the best answer is that for good motor longevity, service factor should not be used for basic load calculations. By not loading the
Centrifugal Pump Types and Applications
113
motor into the service factor, the motor can better withstand adverse conditions that occur. Adverse conditions include higher than normal ambient
temperatures, low or high voltage, voltage imbalances, and occasional overload. These conditions are less likely to damage the motor or shorten its life if
the motor is not loaded into its service factor in normal operation. That being
said, however, there are a good many lighter duty pumps whose motors are
typically sized to run in the service factor. Examples would include light
intermittent duty residential pumps such as utility sump pumps, or small
OEM products where low cost is paramount.
D. Insulation Classes
The electrical portions of every motor must be insulated from contact with
other wires and with the magnetic portion of the motor. The insulation system consists of the varnish that jackets the magnet wire in the windings along
with the slot liners that insulate the wire from the steel laminations. The insulation system also includes tapes, sleeves, tie strings, a final dipping varnish,
and the leads that bring the electrical circuits out to the junction box.
Insulation systems are rated by their resistance to thermal degradation.
The four basic insulation systems normally encountered are Classes A, B,
F, and H. Class A has a temperature rating of 105°C (221°F), and each step
from A to B, B to F, and F to H involves a 25°C (45°F) jump. The insulation
class in any motor must be able to withstand at least the maximum ambient
temperature plus the temperature rise that occurs because of continuous full
load operation. Selecting an insulation class higher than necessary to meet
this minimum can help extend motor life or make a motor more tolerant
of overloads, high ambient temperatures, and other problems that normally
shorten motor life.
A widely used rule of thumb states that every 10°C (18°F) increase in operating temperature cuts insulation life in half. Conversely, a 10°C decrease
doubles insulation life. Choosing a one-step higher insulation class than
required to meet the basic performance specifications of a motor provides
25°C of extra temperature capability. The rule of thumb predicts that this
better insulation system increases the motor’s thermal life expectancy by
approximately 500%.
E. Motor Frame Size
1. Historical Perspective
Industrial electric motors have been widely available for nearly a century,
during which there have been a great many changes. One of the most obvious has been the ability to pack more horsepower in a smaller physical size.
Another important achievement has been the standardization of motors by
the National Electrical Manufacturers Association (NEMA).
—
203
204
224
225
254
284
324
326
364S
364S
365S
404S
405S
444S
445S
504S
505S
—
—
—
182
184
184
213
215
254U
256U
284U
286U
324US
326US
364US
365US
404US
405US
444US
445US
—
—
—
143T
145T
145T
182T
184T
213T
215T
254T
256T
284TS
286TS
324TS
326TS
364TS
365TS
404TS
405TS
444TS
445TS
1964
Rerate
203
204
224
225
254
284
324
326
364
364
365
404
405S
444S
445S
504S
505S
—
—
—
Original
Source: Courtesy of Edward Cowern, P.E., North Haven, CT.
1
1.5
2
3
5
7.5
10
15
20
25
30
40
50
60
75
100
125
150
200
250
1952
Rerate
NEMA
Program HP
Original
3600
RPM
Frame Size Reference Table—Open Drip-Proof Motors
TABLE 4.5
182
184
184
213
215
254U
256U
284U
286U
324U
326U
364U
365US
404US
405US
444US
445US
—
—
—
1952
Rerate
1800
143T
145T
145T
182T
184T
213T
215T
254T
256T
284T
286T
324T
326T
364T
365T
404T
405T
444T
445T
—
1964
Rerate
204
224
225
254
284
324
326
364
365
404
405
444
445
504
505
—
—
—
—
—
Original
184
184
213
215
254U
256U
284U
324U
326U
364U
365U
404U
405U
444U
445U
—
—
—
—
—
1952
Rerate
1200
145T
182T
184T
213T
215T
254T
256T
284T
286T
324T
326T
364T
365T
404T
405T
444T
445T
—
—
—
1964
Rerate
225
254
254
284
324
326
364
365
404
405
444
445
504
505
—
—
—
—
—
—
Original
213
213
215
254U
256U
284U
286U
326U
364U
365U
404U
405U
444U
445U
—
—
—
—
—
—
1952
Rerate
900
182T
184T
213T
215T
254T
256T
284T
286T
324T
326T
364T
365T
404T
405T
444T
445T
—
—
—
—
1964
Rerate
114
Pump Characteristics and Applications
—
203
204
224
225
254
284
324
326
365S
404S
405S
444S
445S
504S
505S
—
—
—
182
184
184
213
215
254U
256U
286U
324U
326US
364US
365US
405US
444US
445US
—
—
—
143T
145T
182T
184T
213T
215T
254T
256T
284TS
286TS
324TS
326TS
364TS
365TS
405TS
444TS
445TS
1964
Rerate
203
204
224
225
254
284
324
326
364
365
404
405
444S
445S
504S
505S
—
—
Original
Source: Courtesy of Edward Cowern, P.E., North Haven, CT.
1
1.5
2
3
5
7.5
10
15
20
25
30
40
50
60
75
100
125
150
1952
Rerate
NEMA
Program HP
Original
3600
RPM
182
184
184
213
215
254U
256U
284U
286U
324U
326U
364U
365US
405US
444US
445US
—
—
1952
Rerate
1800
143T
145T
145T
182T
184T
213T
215T
254T
256T
284T
286T
324T
326T
364T
365T
405T
444T
445T
1964
Rerate
204
224
225
254
284
324
326
364
365
404
405
444
445
504
505
—
—
—
Original
Frame Size Reference Table—Three-Phase Totally Enclosed Fan Cooled Motors
TABLE 4.6
184
184
213
215
254U
256U
284U
324U
326U
364U
365U
404U
405U
444U
445U
—
—
—
1952
Rerate
1200
145T
182T
184T
213T
215T
254T
256T
284T
286T
324T
326T
364T
365T
404T
405T
444T
445T
—
1964
Rerate
225
254
254
284
324
326
364
365
404
405
444
445
504
505
—
—
—
—
Original
213
213
215
254U
256U
284U
286U
326U
364U
365U
404U
405U
444U
445U
—
—
—
—
1952
Rerate
900
182T
184T
213T
215T
254T
256T
284T
286T
324T
326T
364T
365T
404T
405T
444T
445T
—
—
1964
Rerate
Centrifugal Pump Types and Applications
115
116
Pump Characteristics and Applications
A key part of motor interchangeability has been the standardization of
frame sizes. This means that the same horsepower, speed, and enclosure
will normally have the same frame size from different motor manufacturers.
Thus, a motor from one manufacturer can be replaced with a similar motor
from another company provided they are both in standard frame sizes.
The standardization effort over the last 70 years has resulted in one original grouping of frame sizes called “original.” In 1952, new frame assignments were made. These were called “U frames.” The current “T frames”
were introduced in 1964. “T” frames are the current standard and most likely
will continue to be for some time in the future.
Although “T” frames were adopted in 1964, there are still many “U” frame
motors in service that will have to be replaced in the future. Similarly, there
are also many of the original frame size motors (pre-1952) that will reach the
end of their useful life and will have to be replaced. For this reason, it is desirable to have reference material available on frame sizes and some knowledge
of changes that took place as a part of the so-called rerate programs.
Tables 4.5 and 4.6 show the standard frame size assignments for the three
different eras of motors, broken down for open drip-proof (Table 4.5) and
totally enclosed fan cooled (Table 4.6). For each horsepower rating and speed,
there are three different frame sizes. The first is the original frame size, the
middle one is the “U frame” size, and the third one is the “T frame.” These
are handy reference tables because they give general information for all three
vintages of three phase motors in integral horsepower frame sizes.
One important item to remember is that the base mounting hole spacing
(“E” and “F” dimensions) and shaft height (“D” dimension) for all frames
having the same three digits regardless of vintage, will be the same.
2. Rerating and Temperature
The ability to rerate motor frames to get more horsepower in a frame has been
brought about mainly by improvements made in insulating materials. As a
result of this improved insulation, motors can be run much hotter. This allows
more horsepower in a compact frame. For example, the original NEMA frame
sizes ran at very low temperatures. The “U” frame motors were designed for
use with Class A insulation, which has a rating of 105°C. The motor designs
were such that the capability would be used at the hottest spot within the
motor. “T” frame motor designs are based on utilizing Class B insulation with
a temperature rating of 130°C. This increase in temperature capability made it
possible to pack more horsepower into the same size frame. To accommodate
the larger mechanical horsepower capability, shaft and bearing sizes had to
be increased. Thus, you will find that the original 254 frame (5 HP at 1800 rpm)
has a 1 1/8-in shaft. The 254U frame (7 1/2 HP at 1800 rpm) has a 1 ­1/8-in
shaft, and the current 254T frame (15 HP at 1800 rpm) has a 1 3/8-in shaft.
Bearing diameters were also increased to accommodate the larger shaft sizes
and heavier loads associated with the higher horsepowers.
Centrifugal Pump Types and Applications
117
3. Motor Frame Dimensions
Tables 4.7 and 4.8 show the NEMA and IEC motor dimensions, respectively.
Most of the motor dimensions are standard dimensions that are common to
all motor manufacturers. One exception to this is the “C” dimension (overall
motor length), which will change from one manufacturer to another.
4. Fractional Horsepower Motors
The term “fractional horsepower” is used to cover those frame sizes having
two digit designations, as opposed to the three digit designations that are
found in Tables 4.5 and 4.6. The NEMA frame sizes that are normally associated with industrial fractional horsepower motors are 42, 48, and 56. In
this case, each frame size designates a particular shaft height, shaft diameter, and face or base mounting hole pattern. In these motors, specific frame
assignments have not been made by horsepower and speed, so it is possible
that a particular horsepower and speed combination might be found in
three different frame sizes. In this case, for replacement, it is essential that
the frame size be known as well as the horsepower, speed, and enclosure.
The derivation of the two-digit frame number is based on the shaft height
in sixteenths of an inch. A 48-frame motor would have a shaft height of 48
divided by 16, or 3 in. Similarly, a 56-frame motor would have a shaft height
of 3 1/2 inches. The largest of the current fractional horsepower frame sizes
is a 56-frame, which is available in horsepowers greater than those normally associated with fractional motors. For example, 56 frame motors are
built in horsepowers up to 3 HP, and in some cases, 5 HP. For this reason, calling motors with two digit frame sizes “fractionals” is somewhat
misleading.
5. Integral Horsepower Motors
The term “integral horsepower” generally refers to those motors having three
digit frame sizes such as 143T or larger. When dealing with these frame sizes,
one handy rule of thumb is that the centerline shaft height (“D” dimension)
above the bottom of the base is the first two digits of the frame size divided
by four. For example, a 254T frame would have a shaft height of 25 ÷ 4 =
6.25 in. Although the last digit does not directly relate to an “inch” dimension, larger numbers do indicate that the rear bolt holes are moved further
away from the shaft end bolt holes (the “F” dimension becomes larger).
6. Frame Designation Variations
In addition to the standard numbering system for frames, there are some
variations that will appear. These are itemized below along with an explanation of what the various letters represent.
Pump Characteristics and Applications
NEMA Motor Frame Dimensions
TABLE 4.7
118
BA Dim
2-3/4
3-1/2
Source: Courtesy of Baldor Electric Company, Fort Smith, AR.
NEMA C-Face
143-5TC
182-4TC
NEMA C-Face
213-5TC
254-6TC
BA Dim
4-3/4
4-3/4
Centrifugal Pump Types and Applications
119
IEC Motor Frame Dimensions
TABLE 4.8
120
Pump Characteristics and Applications
Source: Courtesy of Baldor Electric Company, Fort Smith, AR.
Centrifugal Pump Types and Applications
121
122
Pump Characteristics and Applications
C This designates a C-face (flange)-mounted motor. This is the most
popular type of face-mounted motor, and is the type used for
close-coupled pumps. The C-face motor has a specific bolt pattern
on the shaft end to allow mounting. The critical items on C-face
motors are the “bolt circle” (AJ dimension), register (also called rabbet) diameter (AK dimension), and the shaft size (U dimension).
C-face motors always have threaded mounting holes in the face of
the motor.
D The “D” flange has a special type of mounting flange installed on
the shaft end. In the case of the “D” flange, the flange diameter is
larger than the body of the motor and it has clearance holes suitable for mounting bolts to pass through from the back of the motor
into threaded holes in the mating part. “D” flange motors are not
as popular as “C” flange motors, and are almost never used on
pumps.
H Used on some 56 frame motors, “H” indicates that the base is suitable for mounting in 56, 143T, or 145T mounting dimensions.
J This designation is used with 56 frame motors and indicates that
the motor is made for jet pump service with a threaded stainless steel
shaft and standard 56C face.
JM The letters “JM” designate a special pump shaft originally designed
for a mechanical seal. This motor also has a C face.
JP Similar to the JM style of motor having a special shaft, the JP motor
was originally designed for a “packing” type of seal. The motor also
has a C face.
S The use of the letter “S” in a motor frame designates that the motor
has a “short shaft.” Short shaft motors have shaft dimensions that are
smaller than the shafts associated with the normal frame size. Short
shaft motors are designed to be directly coupled to a load through a
flexible coupling. They are not supposed to be used on applications
where belts are used to drive the load.
T A “T” at the end of the frame size indicates that the motor is of the
1964 and later “T” frame vintage.
U A “U” at the end of the frame size indicates that the motor falls into
the “U” frame size assignment (1952 to 1964) era.
Y When a “Y” appears as a part of the frame size, it means that the
motor has a special mounting configuration. It is impossible to tell
exactly what the special configuration is, but it does denote that
there is a special nonstandard mounting.
Z Indicates the existence of a special shaft that could be longer, larger,
or have special features such as threads, holes, etc. “Z” indicates only
that the shaft is special in some undefined way.
123
Centrifugal Pump Types and Applications
F. Single-Phase Motors
Three-phase motors start and run in a direction based on the “phase rotation” of the incoming power. Single-phase motors are different. They require
an auxiliary starting means. Once started in a direction, they continue to run
in that direction. Single-phase motors are categorized by the method used to
start the motor and establish the direction of rotation. The three major types
of single-phase motors, shaded pole, split phase, and capacitor motors are summarized in Table 4.9 and are described below.
Shaded pole motors employ the simplest of all single-phase starting methods.
These motors are used only for small, simple applications such as bathroom
exhaust fans. In the shaded pole motor, the motor field poles are notched
and a copper shorting ring is installed around a small section of the poles as
shown in Figure 4.50.
The altered pole configuration delays the magnetic field build-up in
the portion of the poles surrounded by the copper shorting rings. This
arrangement makes the magnetic field around the rotor seem to rotate
from the main pole toward the shaded pole. This appearance of field
TABLE 4.9
Single-Phase Motor Types and Characteristics
Category
Approximate HP Range
Relative Efficiency
Shaded pole
Split phase
Capacitor
1/100 – 1/6 HP
1/25 – 1/2 HP
1/50 – 15 HP
Low
Medium
Medium to high
Copper shorting rings
(delay field build-up)
Notch
Apparent field
rotation (CCW)
Notch
Shaded pole induction motor
FIGURE 4.50
Shaded pole is the simplest of all single-phase starting methods. (Courtesy of Edward Cowern,
P.E., North Haven, CT.)
124
Pump Characteristics and Applications
rotation starts the rotor moving. Once started, the motor accelerates to
full speed.
Split phase motors have two separate windings in the stator (stationary
portion of the motor) (see Figure 4.51). The winding shown in black is only
for starting. It uses a smaller wire size and has higher electrical resistance
than the main winding. The difference in the start winding location and its
altered electrical characteristics causes a delay in current flow between the
two windings. This time delay coupled with the physical location of the starting winding causes the field around the rotor to move and start the motor.
A centrifugal switch or other device disconnects the starting winding when
the motor reaches approximately 75% of rated speed. The motor continues to
run on normal induction motor principles.
Split phase motors are generally available from 1/25 to 1/2 HP. Their main
advantage is low cost. Their disadvantages are low starting torque and high
starting current. These disadvantages generally limit split phase motors to applications where the load needs only low starting torque and starts are infrequent.
Capacitor motors are the most popular single-phase motors on pumps.
They are used in many agricultural, commercial, and industrial applications
where three-phase power is not available. Capacitor motors are available
in sizes from subfractional to 15 HP. Capacitor motors fall into three types,
summarized in Table 4.10, and described in the following paragraphs.
Capacitor start–induction run motors form the largest group of general-purpose single-phase motors. The winding and centrifugal switch arrangement
is similar to that in a split phase motor. However, a capacitor start–induction
Centrifugal switch
(Open)
FIGURE 4.51
Split phase motors have two separate windings in the stator. (Courtesy of Edward Cowern,
P.E., North Haven, CT.)
125
Centrifugal Pump Types and Applications
TABLE 4.10
Capacitor Motor Types and HP Ranges
Category
Usual HP Range
Capacitor start–induction run
Single-value capacitor (permanent split capacitor)
Two-value capacitor (capacitor start–capacitor run)
1/8–3 HP
1/50–1 HP
2–15 HP
run motor has a capacitor in series with the starter winding. Figure 4.52
shows the capacitor start–induction run motor. The starting capacitor produces a time delay between the magnetization of the starting poles and the
running poles, creating the appearance of a rotating field. The rotor starts
moving in the same direction. As the rotor approaches running speed, the
starting switch opens and the motor continues to run in the normal induction motor mode.
This moderately priced motor produces relatively high starting torque
(225%–400% of full load torque) with moderate inrush current. Capacitor
start motors are ideal for hard-to-start loads such as refrigeration compressors. Due to its other desirable characteristics, it is also used in applications
where high starting torque may not be required. The capacitor start motor
can usually be recognized by the bulbous protrusion on the frame that
houses the starting capacitor.
Capacitor
Centrifugal switch
FIGURE 4.52
Capacitor start–induction run motors have a capacitor in series with the starter winding.
(Courtesy of Edward Cowern, P.E., North Haven, CT.)
126
Pump Characteristics and Applications
In some applications, it is not practical to install a centrifugal switch within
the motor. These motors have a relay operated by motor inrush current.
The relay switches the starting capacitor into the circuit during the starting
period. When the motor approaches full speed, the inrush current decreases
and the relay opens to disconnect the starting capacitor.
Single-value capacitor motors, also called permanent split capacitor (PSC)
motors, utilize a capacitor connected in series with one of the two windings.
This type of motor is generally used on small sizes (less than 1 HP). It is ideally suited for small fans, blowers, and pumps. Starting torque on this type
of motor is generally 100%, or less, of full load torque. A PSC motor would
look the same as the capacitor start–induction run motor shown in Figure
4.52 except without the centrifugal switch.
Two-value capacitor motors are utilized in large horsepower (2–15 HP) singlephase motors. Figure 4.53 shows this motor. The running winding, shown
in white, is energized directly from the line. A second winding, shown in
black, serves as a combined starting and running winding. The black winding is energized through two parallel capacitors. Once the motor has started,
a switch disconnects one of the capacitors, letting the motor operate with the
remaining capacitor in series with the winding of the motor.
The two-value capacitor motor starts as a capacitor start motor but runs as
a form of a two-phase or PSC motor. Using this combination, it is possible to
build large single-phase motors having high starting torques and moderate
starting currents at reasonable prices.
Starting
capacitor
Centrifugal switch
Running capacitor
FIGURE 4.53
Two value capacitor motors are used in large horsepower single-phase motors. (Courtesy of
Edward Cowern, P.E., North Haven, CT.)
Centrifugal Pump Types and Applications
127
The two-value capacitor motor frequently uses an oversized conduit box to
house both the starting and running capacitors.
G. Motors Operating on Variable Frequency Drives
Variable frequency drives (VFDs) are discussed in more detail in Chapter 6,
Section IV. In the infancy of variable frequency drives, a major selling point was
that VFDs could adjust the speed of “standard” three-phase induction motors.
This claim was quite true when the variable frequency drives were “six-step”
designs. The claim is still somewhat true, although pulse width modulated
(PWM) VFDs have somewhat changed the rules. PWM drives are electrically
more punishing on motor windings, especially for 460- and 575-V drives.
“Standard” motors can still be used on many VFDs, especially on commercial pump, fan, and blower applications, as long as the motors are highquality, conservative designs that use inverter spike resistant (ISR) magnet
wire. On these variable torque loads, a relatively small speed reduction
results in a dramatic reduction in the torque required from the motor. For
example, a 15% reduction in speed reduces the torque requirement by over
25%, so these motors are not stressed from a thermal point of view. Also,
variable torque loads rarely need a wide speed range. Since the performance
of centrifugal pumps, fans, and blowers falls off dramatically as speed is
reduced, speed reduction below 40% of base speed is rarely required.
The natural question is, “What is meant by a high quality, conservative
design?” Basically, this means that the motor must have phase insulation, should
operate at a relatively low temperature rise (as in the case with most premium
efficiency motors), and should use a high class of insulation (either F or H).
In addition, it is frequently desirable to have a winding thermostat in the motor
that will detect any motor overheat conditions that may occur. Overheating
could result from overload, high ambient temperature, or loss of ventilation.
Inverter-duty motors being offered in the marketplace today incorporate
premium efficiency designs along with oversized frames or external blowers to
cool the motor regardless of its speed. These motors are primarily designed
for constant torque loads where the affinity laws do not apply. Inverter-duty
motors usually have winding thermostats that shut down the motor through
the VFD control circuit in case of elevated temperature inside the motor.
Inverter-duty motors also have high-temperature insulating materials operated at lower temperatures. This reduces the stress on the insulation system.
Although some of the design features of inverter-duty motors are desirable
for centrifugal pump applications using VFDs, these applications usually do
not require inverter-duty motors, which typically cost a good deal more than
regular premium efficiency motors.
Some caution should be observed. Generally speaking, the power coming
out of a VFD is somewhat rougher on the motor than power from a pure
60-cycle source. Thus, it is not a good idea to operate motors on VFDs into
their service factors.
128
Pump Characteristics and Applications
In addition, when an old motor (one that has been in service for some time)
is to be repowered from a variable frequency drive, it may be desirable to
add a load reactor between the VFD and the motor. The reactor reduces the
stress on the motor windings by smoothing out current variations, thereby
prolonging motor life.
Reactors are similar to transformers with copper coils wound around a
magnetic core. Load reactors increase in importance when the VFDs are
going to run in the “quiet” mode. In this mode, the very high carrier frequency can create standing waves that potentially double the voltage peaks
applied to the motor. The higher voltage can stress the motor insulation
enough to cause premature failure.
H. NEMA Locked Rotor Code
The “NEMA code letter” is an additional piece of information on the motor
nameplate. These letters indicate a range of inrush (starting or locked rotor)
currents that occur when a motor starts across the line with a standard magnetic or manual starter. Most motors draw 5 to 7 times rated full load (nameplate) amps during the time it takes to go from standstill up to about 80% of
full-load speed. The length of time the inrush current lasts depends on the
amount of inertia (flywheel effect) in the load. On centrifugal pumps with
very low inertia, the inrush current lasts only a few seconds. On large, squirrel cage blowers, the inrush current can last considerably longer.
The locked rotor code letter quantifies the value of the inrush current for a
specific motor. The lower the code letter, the lower the inrush current. Higher
code letters indicate higher inrush currents. Table 4.11 lists the NEMA locked
rotor code letters and their parameters.
TABLE 4.11
NEMA Locked Rotor Code Letters and Their Parameters
NEMA Code
Letter
A
B
C
D
E
F
G
H
I
J
K
Locked Rotor
KVA/HP
NEMA
Code Letter
Locked Rotor
KVA/HP
0–3.15
3.15–3.55
3.55–4.0
4.0–4.5
4.5–5.0
5.0–5.6
5.6–6.3
6.3–7.1
Not used
7.1–8.0
8.0–9.0
L
M
N
O
P
Q
R
S
T
U
V
9.0–10.0
10.0–11.2
11.2–12.5
Not used
12.5–14.0
Not used
14.0–16.0
16.0–18.0
18.0–20.0
20.0–22.4
22.4 and up
Centrifugal Pump Types and Applications
129
I. Amps, Watts, Power Factor, and Efficiency
1. Introduction
There seems to be a great deal of confusion among the users of electric motors
regarding the relative importance of power factor, efficiency, and amperage,
particularly as related to operating cost. The following information should
help to put these terms into proper perspective.
At the risk of treating these items in reverse order, it might be helpful to
understand that in an electric bill, commercial, industrial, or residential,
the basic unit of measurement is the kilowatt-hour. This is a measure of
the amount of energy that is delivered. In many respects, the kilowatt-hour
could be compared with a ton of coal, a cubic foot of natural gas, or a gallon of gasoline, in that it is a basic energy unit. The kilowatt-hour is not
directly related to amperes, and at no place on an electric bill will you find
any reference to the amperes that have been utilized. It is vitally important
to note this distinction. The user is billed for kilowatt-hours, not necessarily
for amperes.
2. Power Factor
Perhaps the greatest confusion arises due to the fact that early in our science
educations, we were told that the formula for watts was amps times volts.
This formula, watts = amps × volts, is perfectly true for direct current circuits. It also works on some AC loads such as incandescent light bulbs, quartz
heaters, electric range heating elements, and other equipment of this general
nature. However, when the loads involve a characteristic called inductance,
the formula has to be altered to include a new term called power factor. Thus,
the new formula for single-phase loads becomes watts are equal to amps ×
volts × power factor. The new term, power factor, is always involved in applications where AC power is used and inductive magnetic elements exist in
the circuit. Inductive elements are magnetic devices such as solenoid coils,
motor windings, transformer windings, fluorescent lamp ballasts, and similar equipment that have magnetic components as part of their design.
Looking at the electrical flow into this type of device, we find that there
are, in essence, two components. One portion is absorbed and utilized to
do useful work. This portion is called the real power. The second portion
is literally borrowed from the power company and used to magnetize the
magnetic portion of the circuit. Due to the reversing nature of AC power,
this borrowed power is subsequently returned to the power system when
the AC cycle reverses. This borrowing and returning occurs on a continuous
basis. Power factor then becomes a measurement of the amount of real power
that is used, divided by the total amount of power, both borrowed and used.
Values for power factor will range from zero to 1.0. If all the power is borrowed and returned with none being used, the power factor would be zero.
If on the other hand, all of the power drawn from the power line is utilized
130
Pump Characteristics and Applications
Typical characteristics
60
12,000
8000
4000
80
60
50
Efficienc
40
Amp
Amps
Watts
16,000
% Efficiency and power factor
100
15 HP, 1725, 3 , 60 Hz
200 V
30
40
20
20
10
0
0
Pow
y
s
er f
act
or
Rated load
tts
Wa
0
2
4
6
8
10
Horsepower
12
14
16
18
FIGURE 4.54
Typical motor performance data. (Courtesy of Edward Cowern, P.E., North Haven, CT.)
and none is returned, the power factor becomes 1.0. In the case of electric
heating elements, incandescent light bulbs, etc., the power factor is 1.0. In
the case of electric motors, the power factor is variable and changes with the
amount of load that is applied to the motor. Thus, a motor running on a work
bench, with no load applied to the shaft, will have a low power factor (perhaps 0.1 or 10%), and a motor running at full load, connected to a pump or a
fan might have a relatively high power factor (perhaps 0.78 or 78%). Between
the no-load point and the full-load point, the power factor increases steadily
with the horsepower loading that is applied to the motor. These trends can
be seen on the typical motor performance data plots that are shown for a
15-HP, 1725-rpm, three-phase motor in Figure 4.54.
3. Efficiency
Now, let us consider one of the most critical elements involved in motor operating cost, motor efficiency. Efficiency is the measure of how well the electric
motor converts the power that is purchased into useful work. For example,
an electric heater such as the element in an electric stove converts 100% of the
power delivered into heat. In other devices such as motors, not all of the purchased energy is converted into usable energy. A certain portion is lost and is
not recoverable because it is expended in the losses associated with operating
the device. In an electric motor, these typical losses are the copper losses, the
iron losses, and the so-called friction and windage losses associated with spinning the rotor and the bearings and moving the cooling air through the motor.
Centrifugal Pump Types and Applications
131
In an energy-efficient (premium efficiency) motor, the losses are reduced
using designs that employ better grades of material, more material, and better designs to minimize the various items that contribute to the losses in the
motor. For example, on a 10-HP motor, an energy-efficient design might have
a full-load efficiency of 91.7%, meaning that, at full load (10 HP), it converts
91.7% of the energy it receives into useful work. A less efficient motor might
have an efficiency of 82%, which would indicate that it only converts 82% of
the power into useful work.
In general, larger motors can be expected to be found with higher efficiencies than smaller ones. For example, a 1-HP, 1800-rpm, premium-efficiency motor might have an efficiency of 85.5%, whereas a 100-HP, 1800-rpm,
premium-efficiency motor might have an efficiency of 95.5%. In general, the
efficiency of most motors will be relatively constant from 50% to 100% of
rated load, although there may be a few percentage points fluctuation in this
range.
4. Amperes
Now let us discuss amperes (abbreviated A). Amperes are an indication of
the flow of electric current into the motor. This flow includes both the borrowed as well as the used power. At low load levels, the borrowed power
is a high percentage of the total power. As the load increases on the motor,
the borrowed power becomes less and less of a factor and the used power
becomes greater. Thus, there is an increase in the power factor as the load
on the motor increases. As the load continues to increase beyond 50% of the
rating of the motor, the amperage starts to increase in a nearly straight-line
relationship. This can be seen in Figure 4.54.
5. Summary
Figure 4.54 shows significant items that have been discussed as plots of efficiency, power factor, amps, and watts, as they relate to horsepower. The most
significant factor of all these is the watts requirement of the motor for the
various load levels, because it is the watts that will determine the operating
cost of the motor, not the amperage.
The user having an extremely low power factor in the total plant electrical
system may be penalized by the utility company because the user is effectively borrowing a great deal of power without paying for it. When this type
of charge is levied on the customer, it is generally called a power factor penalty.
In general, power factor penalties are levied only on large industrial customers and rarely on smaller customers, regardless of their power factor. In addition, there are a great many types of power customers such as commercial
establishments, hospitals, and some industrial plants that inherently run at
very high power factors. Thus, the power factor of individual small motors
132
Pump Characteristics and Applications
that are added to the system will not have any significant effect on the total
plant power factor.
It is for these reasons that the blanket statement can be made that increasing motor efficiency will reduce the kilowatt-hour consumption and the
power cost for all classes of power users, regardless of their particular rate
structure or power factor situation. This same type of statement cannot be
made relative to power factor.
The following formulas are useful for calculating operating costs for electric motors.
kilowatt-hours = (HP × 0.746 × hours of operation)/motor efficiency (4.2)
where HP is the average-load HP (which may be lower than motor nameplate HP), and motor efficiency is expressed as a decimal.
kilowatt-hours = (watts × hours of operation)/1000
(4.3)
5
Sealing Systems and Sealless Pumps
I. Overview
This chapter addresses sealing in pumps. The chapter begins with a discussion of O-rings, which are widely used as sealing elements in pumps.
The chapter then moves on to shaft sealing. Most pumps require sealing
of the shaft where it penetrates the pump casing. The casing at the point of
shaft penetration is subjected to either a positive pressure or a vacuum. Some
pump types such as horizontal split case are sealed at both ends, while other
types such as end suction have only one sealing point.
A very basic method for shaft sealing is the packed stuffing box. The various components of the stuffing box and packing assembly are explained,
along with limitations on their use.
Mechanical shaft seals are described as to their benefits, general function,
description, and application. Particular types of mechanical seals (inside vs.
outside, unbalanced vs. balanced, single vs. dual) are explained and illustrated. A relatively new type of noncontacting, gas-lubricated seal is introduced in this chapter.
Finally, several types of centrifugal pumps that require no packing or
seal are introduced. Sealless pumping is an important technology not only
because it guarantees zero emissions, but also because it eliminates the seal,
a component that may require frequent maintenance. The two major types
of sealless centrifugals, magnetic drive and canned motor pumps, are described
and compared.
II. O-Rings
For those readers who are unfamiliar with O-ring sealing technology, this
section should aid in the understanding of the O-ring, its basic design principles, and the many sealing functions an O-ring can be expected to perform
when properly specified and installed. The following material on O-rings
133
134
Pump Characteristics and Applications
is printed with permission of the Parker Hannifin Corporation, Seal Group.
See also discussion of the various O-ring materials in Chapter 7, Section VI.
A. What Is an O-Ring?
An O-ring (Figure 5.1) is a torus, or a doughnut-shaped object, generally
made from an elastomer, although such materials as plastics and metal are
sometimes used. This section will deal entirely with elastomeric O-rings
used for sealing purposes.
B. Basic Principle of the O-Ring Seal
An O-ring seal is a means of closing off a passageway and preventing an
unwanted loss or transfer of fluid. The classic O-ring seal consists of two
elements, the O-ring itself, and a properly designed gland or cavity to contain the elastomeric material. Prevention of the fluid loss or transfer may be
obtained by several methods: welding, soldering, brazing, or the yielding of
a softer material wholly or partially confined between two mating surfaces.
This latter method best describes the design principle behind the operation
of an O-ring seal.
C. The Function of the O-Ring
The elastomer is contained in the gland and forced to flow into the surface imperfections of the glands and any clearance available to it, creating a condition of “zero” clearance and thus effecting a positive block to
the fluid being sealed (Figure 5.2). The pressure that forces the O-ring to
flow is supplied by mechanical pressure, or “squeeze,” generated by proper
gland design and material selection, and by system pressure transmitted by the fluid itself to the seal element. In fact, the classic O-ring seal
I.D.
Cross section or width
An elastomeric
O-ring section
FIGURE 5.1
Anatomy of an O-ring. (Courtesy of Parker Hannifin Corporation, Seal Group.)
Sealing Systems and Sealless Pumps
135
FIGURE 5.2
O-ring seal installed under minimal pressure. (Courtesy of Parker Hannifin Corporation, Seal
Group.)
may be said to be “pressure assisted” (Figure 5.3) in that the more system
pressure, the more effective the seal, until the physical limits of the elastomer are exceeded and the O-ring begins to extrude into the clearance gap
(Figure 5.4). This condition can usually be avoided by proper gland design
and material selection.
FIGURE 5.3
O-ring seal under increasing pressure from the left. (Courtesy of Parker Hannifin Corporation,
Seal Group.)
FIGURE 5.4
O-ring seal extruding with pressure limit exceeded. (Courtesy of Parker Hannifin Corporation,
Seal Group.)
136
Pump Characteristics and Applications
D. Static and Dynamic O-Ring Sealing Applications
O-ring seals are generally divided into two main groups, static seals, in which
there is little or no relative motion between the mating surfaces, and dynamic
seals, which must function between surfaces with definite relative motion,
such as the seal on the piston of a hydraulic cylinder. Of the two types,
dynamic sealing is the more difficult and requires more critical design work
and materials selection.
The most common type of dynamic motion utilizing an O-ring sealing
system is reciprocal motion as found in hydraulic cylinders, actuators, and the
like.
E. Other Common O-Ring Seal Configurations
Aside from reciprocating seals, there are other types of motion where an
O-ring seal may be utilized. For example,
Oscillating seals (Figure 5.5), in which the inner or outer member of the
assembly moves in an arc relative to the other, rotating one of the members
in relation to the O-ring.
Rotary seals (Figure 5.6), where an inner or outer member of the sealing
assembly revolves around the shaft axis in only one direction. The direction may be reversed. If there are multiple brief arcs of motion, the designer
should refer to parameters for oscillating seals.
Seat seals (Figure 5.7) utilize an O-ring to close a flow passage by distorting
the face of the O-ring against the opposite contact face. Closing the passage
distorts the seal element to create the closure.
Crush seals (Figure 5.8), a variation of the static seal, literally “crush” an
O-ring into a space with a cross section different from the standard gland.
Although often an effective seal, the O-ring is permanently deformed and
must be replaced if the unit is opened.
Pneumatic seals may be any of the previously described types, but are given
a different classification because they seal a gas or vapor as opposed to a
liquid. Thus, other design factors such as adequate lubrication (for dynamic
seals), temperature increases due to compression of gases, and permeability
of the seal element must be considered when the application is pneumatic.
FIGURE 5.5
Oscillating seal. (Courtesy of Parker Hannifin Corporation, Seal Group.)
137
Sealing Systems and Sealless Pumps
Note that groove
size prevents
rotation of O-ring
FIGURE 5.6
Rotary seal. (Courtesy of Parker Hannifin Corporation, Seal Group.)
FIGURE 5.7
Seat seal. (Courtesy of Parker Hannifin Corporation, Seal Group.)
O-ring volume is usually
90%–95% gland volume
FIGURE 5.8
Crush seal. (Courtesy of Parker Hannifin Corporation, Seal Group.)
Vacuum seals also may be any of the foregoing types (except pneumatic),
and are classified separately because of special design considerations and
the unusually low leakage requirements of vacuum systems.
F. Limitations of O-Ring Use
Although O-rings offer a dependable and reasonably economical approach
to hydraulic sealing, they are not a “universal” seal, applicable to all sealing
138
Pump Characteristics and Applications
problems. Certain limitations must be imposed upon their use; among them,
high temperatures, high frictional (rubbing) speeds, cylinder ports over
which the seal element must pass, and excessive clearances. O-rings therefore may be considered for just about all sealing applications except
• Rotary speeds exceeding 1500-ft/min contact speed
• Environments (temperature and media) incompatible with any elastomeric material
• Insufficient structural support for anything but a flat gasket
III. Stuffing Box and Packing Assembly
Figure 5.9 shows a stuffing box and packing assembly, the oldest and one of
the most common shaft sealing systems for pumps. In Figure 5.9, the inside
of the pump casing is on the left side of the figure, and the environment is
on the right side. There are five components of the stuffing box and packing
assembly, as described below.
Lantern
ring
Sealing liquid
connection
Packing gland
(quench type)
Stuffing
box
bushing
Impeller
Stuffing
box
throat
Mechanical
packing
Stuffing
box
FIGURE 5.9
Stuffing box and packing assembly. (Courtesy of Goulds Pumps Inc., a subsidiary of ITT
Corporation.)
Sealing Systems and Sealless Pumps
139
A. Stuffing Box
The stuffing box houses the packing assembly and is the location where the
shaft penetrates the casing that is under pressure or vacuum. In some pump
configurations, the stuffing box is actually part of the pump casing, whereas
for other pump types, the stuffing box is bolted to the casing. The stuffing
box is machined so that the throat bore is concentric with the shaft. Its bore
size and depth are designed to accommodate a specific size of packing and
number of packing rings. Many stuffing boxes are fitted with drilled connections that allow the introduction of liquid into the packing area, or provide a
path for liquid to escape to reduce the pressure on the packing, as discussed
in Section III.E.
B. Stuffing Box Bushing
The stuffing box bushing is located at the bottom of the stuffing box, below
the first ring of packing. It serves several functions. The bushing has a close
running clearance between itself and the shaft or sleeve. This allows some
amount of pressure breakdown as liquid throttles across it. In general, the
lower the pressure against which the packing is sealing, the longer the
expected service life of the packing. The bushing also serves to keep larger
solids from entering the packing area where they might embed in the packing and cause undue wear on the shaft or sleeve. The bushing is replaceable
so that the close clearance can be maintained throughout the pump’s life.
The stuffing box bushing keeps wear from occurring in the stuffing box
itself, a much more expensive component to replace than the bushing.
The stuffing box bushing serves as a landing or shoulder for the lowest
ring of packing. If the clearance here is larger than provided by the stuffing
box bushing, the soft, pliable packing could extrude into the clearance as the
packing is tightened down by the gland (see Section III.D below), reducing
the effectiveness of the packing and making its removal more difficult.
The stuffing box bushing also restricts the amount of liquid that might
leak out of the stuffing box in the event of a complete failure of the packing
rings and gland.
Inexpensive lighter-duty pumps may not be supplied with a stuffing box
bushing, but it is a good idea to have such a bushing on any packed pump if
one is available, for the reasons just mentioned.
C. Packing Rings
The mechanical packing rings themselves constitute the heart of the sealing
device. Packing rings are usually formed of a pliable braided fibrous material,
often impregnated with a lubricating medium. The packing material is square
in cross section and is formed into rings that are compressed into the stuffing
box bore and around the shaft. These rings provide a static seal between their
140
Pump Characteristics and Applications
outside diameter and the stuffing box (where there is no relative movement
when the pump is operating) so that no liquid can leak out this way.
The packing rings also compress against the rotating shaft or sleeve, causing the liquid to break down in pressure as it leaks between the packing and
the shaft or sleeve. To operate properly, packing must leak. As the liquid leaks
between the packing and the shaft or sleeve, it cools and lubricates this interface. A common rule of thumb is that a packing assembly should be allowed to
leak around 60 drops per minute to provide adequate lubrication and cooling
of the packing and the surface against which it is running. This is not always
properly understood by equipment owners or maintenance technicians, who
sometimes keep tightening up on the packing gland to minimize the “mess”
of the leaking packing assembly. If too little liquid is allowed to leak across
the packing area, the packing can overheat due to the friction between itself
and the shaft or sleeve. This can cause the packing to degrade quickly, and can
cause wear on the shaft or sleeve. Both of these degradations, in turn, cause
the packing area to begin to leak excessively. It should be noted that there are
several manufacturers that offer drip-less, dry running, self-lubricated packing; however, the author has limited experience with this packing.
Most packed pumps have from four to seven rings of packing, depending
on the amount of pressure to be sealed. Packing cross-sectional width varies
in size, depending on the size of the pump, with common width sizes varying between 1/4 and 3/4 in. Packing can be purchased in preformed rings,
made to fit specific shaft/sleeve diameters and specific stuffing box bores. It
can also be purchased in rolls and cut to the proper length as it is installed.
If packing is properly chosen to be compatible with the liquid pumped,
is properly installed, and is allowed to be lubricated by the pumped liquid,
it can seal relatively high pressures with only periodic replacement. It does
leak, however, as indicated above, and thus is not acceptable for many corrosive or extremely high-pressure services.
If the liquid being pumped contains abrasives, lubricating the packing with
an external flush medium is a good idea. Otherwise, the abrasive particles can
embed in the soft packing and act as a grinding wheel against the shaft or sleeve,
damaging the shaft or sleeve, and causing the packing to leak excessively. See
further discussion of external flushing in Section III.E on lantern rings. If external flushing with a clean liquid is not assured for an abrasive application, the
shaft or sleeve should be coated with a ceramic or other hard coating where it
runs under the packing to protect the shaft or sleeve against early failure.
Refer to Chapter 8, Section IV.A.2, for further guidelines on the proper
installation of packing.
D. Packing Gland
The gland, sometimes called the packing follower, holds the packing rings in
place in the stuffing box and provides a means of further compressing the
Sealing Systems and Sealless Pumps
141
packing as wear occurs on the packing, shaft, or sleeve. Most glands are split,
so that they can be removed to allow more room for changing the packing
without disassembling the shaft coupling. The gland typically attaches to
the stuffing box with studs, and tightening the stud nuts moves the gland
toward the packing, tightening it further.
Some glands have a hollow cavity that allows the introduction of cooling
liquid if the pumped liquid is hot. One important application for this gland
quench feature is for boiler feed service, to prevent the water from flashing
back to steam as it breaks down in pressure across the packing.
E. Lantern Ring
The lantern ring has the same general shape as a ring of packing, fitting
into the stuffing box between packing rings. The lantern ring is typically
made of bronze, stainless steel, or plastic, and is drilled with radial holes
and provided with a channel on its outside and inside perimeter. The lantern ring is fitted into the stuffing box at the location of a drilled opening in the stuffing box. The lantern ring in this case has several possible
functions. Located at roughly the middle of the set of packing rings, the
lantern ring allows the introduction of liquid to lubricate the packing if
the pumped liquid is unable to do this. For example, if the packing is used
on a split case double suction pump (Figure 4.16) operating on a lift, the
pressure just below the packing is below atmospheric. Therefore, the liquid does not tend to leak across the packing area as it would if the packing were sealing against a positive pressure. The usual arrangement is to
run tubing from the pump discharge nozzle or from the high-pressure
part of the casing around to the stuffing box, introducing high-pressure
liquid into the lantern ring. This liquid serves the dual purpose of both
lubricating the packing and sealing against air leaking into the suction of
the pump.
The lantern ring can also be used to introduce an external lubricating
medium if the liquid being pumped has poor lubricity.
If the pumped liquid contains abrasives, the packing can be lubricated by
introducing an external clean flush liquid into the lantern ring. In some cases
with abrasive liquids, the lantern ring and the drilled opening in the stuffing
box are located at the bottom of the stuffing box rather than in the middle of
the packing rings, to more completely eliminate the possibility of solids getting into the packing area.
Another use of a lantern ring is with high-pressure applications. For these
applications, the lantern ring is often located at the bottom of the stuffing
box, just above the stuffing box bushing, but again in line with a drilled hole
in the stuffing box. The lantern ring allows high-pressure liquid to relieve
back to suction, thus lowering the pressure against which the packing rings
have to seal.
142
Pump Characteristics and Applications
IV. Mechanical Seals
A. Mechanical Seal Advantages
Mechanical seals have a number of advantages over stuffing box and packing arrangements. The most important advantages of mechanical seals are
discussed below.
1. Lower Mechanical Losses
In Chapter 2, Section V, the mechanical losses are listed as one of the four factors causing a pump to be inefficient. One of the largest of these mechanical
losses comes from the frictional drag of the shaft running against packing.
A mechanical seal, on the other hand, has considerably less mechanical loss,
thus improving pump efficiency. Some seals eliminate virtually all of the
mechanical friction.
2. Less Sleeve Wear
Section III.C above discusses the fact that the sleeve under the packing is
subject to wear, especially if the pumped liquid has any abrasives in it or if
the packing is tightened too much. With a mechanical seal, this wear of the
shaft sleeve is eliminated.
3. Zero or Minimal Leakage
Whereas packing should leak around 60 drops per minute to be properly
lubricated, the leakage from most mechanical seals is very nearly zero under
normal operation. This is especially important when the pumped liquid is
corrosive, volatile, toxic, or radioactive. Mechanical seal technology is continuously changing to ensure compliance with the emission requirements of
the EPA and other regulatory agencies.
4. Reduced Maintenance
If (and these are big “ifs”) the seal is properly selected for the application,
if the pump is properly aligned and balanced and not subjected to vibrations from other sources and if the seal is properly flushed as appropriate,
a mechanical seal should require less periodic maintenance than a packed
stuffing box. In general, pumps with mechanical seals are less forgiving of
misalignment, imbalance, and vibration than packed pumps.
5. Seal Higher Pressures
Mechanical seals are able to seal against higher pressures and can be used
with higher-speed pumps than packing.
143
Sealing Systems and Sealless Pumps
B. How Mechanical Seals Work
There are many different types of mechanical seals. Understanding seal
types is complicated by the fact that mechanical seal and pump suppliers do
not all use the same terminology to describe the seal components. Seal types
that this chapter describes include inside and outside seals, balanced and
unbalanced seals, and single, and dual seals.
Every mechanical seal has in common at least three sealing points, consisting of two static seals and one dynamic seal. The term “static” seal means
that there is no relative motion between the two parts. Each mechanical seal
has a static seal between the rotating assembly and the sleeve (or shaft if
there is no sleeve) and a second static seal between the stationary part of the
seal and the gland or seal housing. The dynamic seal is between the two seal
faces, one rotating and the other stationary. The descriptions of the different
sealing types to follow serve to clarify these definitions.
Figures 5.10 and 5.11 show basic single mechanical seals, with the liquid
being sealed coming from the left side of the figure, and the shaft penetration
to atmosphere being located on the right side. The part of the seal that is normally rotating is called the seal head. The seal head includes a metal retainer,
typically locked to the sleeve or shaft with setscrews, a spring assembly, and
the seal face (also called the primary ring, washer, seal ring face, or seal ring).
Note that there are variations of seal designs where the seal head is stationary, such as shown in Figure 5.11.
The primary ring runs against the mating ring, or seat. In most arrangements, the mating ring is fixed to the seal housing or gland (such as the case
in Figure 5.10). Although less common, there are designs that have the mating ring rotating, as in Figure 5.11. In the case of the seals shown in Figures
Gland
plate assembly
Socket head
cap screw
Spacer
Gasket
Retainer
Primary O-ring Set
Sleeve
Spring
screw
ring
O-ring O-ring
O-ring Mating
Snap ring
ring
Collar
FIGURE 5.10
Basic single mechanical seal. (Courtesy of John Crane Inc., Morton Grove, IL.)
144
Pump Characteristics and Applications
FIGURE 5.11
Single, inside, balanced seal with stationary seal head and rotating mating ring. (Courtesy of
John Crane Inc., Morton Grove, IL.)
5.10 and 5.11, the mating ring is held in place by O-rings. Other designs have
the mating ring clamped between the gland and the seal housing.
The coil springs shown in Figure 5.10 and the wave spring shown in Figure
5.11 put tension on the primary ring to keep the dynamic seal faces in uniform contact with each other even if the shaft moves axially due to thermal expansion or small amounts of misalignment. If the seal is working
against a positive pressure, the pressure acting against the rotating seal face
also keeps the two seal faces in contact. However, if the liquid being sealed
against is below atmospheric pressure (as is the case for seals on a double
suction pump operating on a suction lift), then the liquid being sealed does
not push the two seal faces together. In that case, the springs must hold the
two faces in contact. To maintain proper spring tension so that the two seal
faces are always in contact, the location of the rotating part of the seal with
respect to the sleeve or shaft (the distance to the edge of the sleeve or shaft)
is an important dimension that the seal and/or pump manufacturer should
provide. Ensuring the proper location of the seal on the shaft is the purpose
Sealing Systems and Sealless Pumps
145
of the spacer and its associated socket head capscrews shown on the seals
in Figures 5.10 and 5.11. This spacer would be removed after the seal is set,
before starting up the pump.
On the seal shown in Figure 5.10, the static seals between the primary
ring and the retainer, and between the sleeve and the shaft are O-rings. This
keeps liquid from leaking along the sleeve to atmosphere. In some seal types,
the static seal between the head and the sleeve is a rubber boot or bellows,
located inside the head. The static seal between the mating ring and the
gland is also an O-ring, which holds and seals the stationary mating ring.
Finally, the dynamic seal takes place at the contact area between the rotating
and stationary seal faces.
The seal faces must be manufactured to extremely tight tolerances to make
sure the seal faces are completely flat. The seal faces are polished, by a process called lapping, so precisely that their flatness is measured in light bands
and their surface finish in microinches. Care should be taken when handling
mechanical seal faces. Even oily fingerprints can be enough to keep the seal
faces from being in perfect contact with each other.
The spring assembly in the mechanical seal may consist of a number of
small coil springs evenly spaced around the periphery, a larger single coil
spring wrapped around the sleeve, or a wave spring. The multiple coil and
wave spring arrangements provide a more uniform contact of the seal faces.
However, the large single-coil spring is stronger and less likely to fail from
corrosion or to fail to perform if the pumped liquid contains solids or viscous
material. There are alternative designs that use various types of metallic or
elastomeric bellows or wave springs in the head assembly.
Just as packing must be lubricated to keep from overheating and causing
the pump to fail, so must the dynamic seal faces of most mechanical seals be
wetted to work properly. This is true even if one of the seal faces is made of
carbon, a common seal face material that has good self-lubricating properties. Because of the very close contact of the mechanical seal faces, the liquid
breaks down in pressure as it passes across the seal face, so that very little, if
any, liquid actually escapes the seal.
In some applications, if the liquid being sealed against is close to its vapor
pressure, there is the concern that the liquid might drop below the vapor
pressure as it leaks across the seal faces. This could cause the dynamic seal
faces not to be wetted across their entire surface, in turn causing them to run
hot. The heat generated could cause the point along the dynamic seal faces at
which the liquid flashes to vapor to move outward progressively, eventually
causing the seal faces to run completely dry and fail prematurely. This must
be prevented from occurring, and is usually done by injecting seal flush liquid from the pump discharge or from an external source into the seal chamber to increase the pressure of the liquid at the seal faces, so that it cannot
drop below its vapor pressure prior to passing across the seal contact faces.
Mechanical seal materials must be carefully chosen to meet the requirements of the application. The metal parts of the seal must be able to handle
146
Pump Characteristics and Applications
the corrosive nature of the liquid. The elastomers used for the static seals
(O-rings, wedges, bellows, and gaskets) must be able to handle the temperature, pressure, and corrosiveness of the liquid. Elastomeric materials used for
these components in mechanical seals include butyl, EPR, Buna-N, neoprene,
VITON® fluoroelastomer, and Teflon® PTFE. (See Chapter 7, Section VI, for a
discussion of O-ring materials.)
The rotating and stationary dynamic seal faces should be constructed of
the best-suited materials for the application. Various grades of carbon are
often used for one of the dynamic sealing faces. Other seal face materials
used for certain applications include ceramic, Ni-resist, tungsten carbide,
and silicon carbide.
To allow the seal and/or pump supplier to assist in making the optimum
selection of seal type and material, it is important to provide as much information as possible about the conditions of service. This includes detailed
information about the liquid being pumped (liquid type, pressure at the seal
housing, temperature, pH, viscosity, vapor pressure, and solids concentration), both at normal operation and the expected range. Also, information
about the type of operation (continuous vs. intermittent, pump speed, alignment, and vibration tolerances) helps to select the best seal for the application.
Finally, the dimensional data related to the pump sleeve outside diameter
and depth of seal housing ensure that a correct seal size are chosen.
If a factory performance test is specified for a pump with a mechanical seal,
the user may want to consider having the pump tested without the mechanical seal installed (substituting a laboratory seal or packing). Operation of
the pump in the manufacturer’s test facility may actually be more severe
service than the pump can expect in the field, if the pump’s service is a clean
liquid with good lubricating properties. Consequently, the seal faces may
not have been selected to work well in the environment of the test facility,
and they may wear excessively during the test. As a compromise measure
in this regard, the user’s seal can be installed for the test, and then the faces
re-lapped following the test.
C. Types of Mechanical Seals
1. Single, Inside Seals
Figures 5.10, 5.11, 5.12, and 5.13 show single, inside mechanical seals. (In these
figures, the sealed liquid is on the left side and atmosphere is on the right
side.) They are called single seals because there is one dynamic seal face and
are called inside seals because the seal parts are located inside the seal housing or gland, unobservable when the pump is running. The single, inside
seal is the most common type of mechanical seal for pumps.
Note that the gland or seal housing in many seals contains a port for injecting liquid, either from the pump discharge or from an external source. This
injection could be used to flush the dynamic seal faces. It could also be used
147
Sealing Systems and Sealless Pumps
Seal flush
Process fluid
Atmosphere
FIGURE 5.12
Single, inside, unbalanced seal. (Courtesy of John Crane Inc., Morton Grove, IL.)
to increase the pressure of the liquid at the seal faces to keep it from dropping
below the vapor pressure of the liquid as described in Section IV.B above. If
the pumped liquid is used for injection, and if the pumped liquid contains
abrasives or other solids, a separator or filter should be used to supply the
seal with clean liquid.
Seal flush
Vent and drain
FIGURE 5.13
Single, inside, balanced seal. (Courtesy of John Crane Inc., Morton Grove, IL.)
148
Pump Characteristics and Applications
Seal flush
FIGURE 5.14
Single, outside, balanced seal. (Courtesy of John Crane Inc., Morton Grove, IL.)
Another important gland feature is a vent-and-drain connection. This feature is shown on the seal in Figure 5.13 and can be supplied as an option with
most mechanical seals. (It may not be available in very small pumps due to
space limitations in the seal area.) The function of the vent and drain connection is to carry away any liquid that leaks past the dynamic seal faces to keep
this liquid from moving along the shaft and then being slung off the shaft.
Depending on the nature of the pumped liquid, failure to drain this liquid
with a vent and drain connection could cause damage to the equipment or be
a safety hazard, in addition to making a mess. Therefore, the vent and drain
feature has a housekeeping function, serves to reduce maintenance expense,
and may even have a safety function. Note that this is an optional feature
for many pumps, and will not necessarily be supplied unless it is specified.
It is a particularly worthwhile feature to include in any seal installation if it
is available.
If the pumped liquid is volatile, there should be a nonsparking bushing
located in the gland in the event the shaft comes in contact with the gland.
The seal in Figure 5.14 shows such a bushing.
2. Single, Outside Seals
Figure 5.14 shows a single, outside mounted seal. Here, the entire rotating
portion of the seal is outside the seal chamber. (In this figure, the sealed liquid is on the left side, atmosphere on the right side.) The static seal between
the primary ring and the shaft is an O-ring, and the static seal between the
mating ring and the gland is also an O-ring.
Sealing Systems and Sealless Pumps
149
The main advantage of an outside seal is that the hardware items of the
seal do not come in contact with the liquid, so this assembly is not subject to
corrosive attack or other deteriorating influences from the pumped liquid.
Also, the rotating assembly is observable during operation, and may be considered easier to get at for maintenance.
The primary disadvantage of the outside seal is that the pressure of the
sealed liquid pushes the two seal faces apart, rather than forcing them
together as is the case for an inside seal. The only thing holding the two seal
faces together is the spring compression. Outside seals are limited to sealing pressures of about 25 to 30 psig unbalanced and about 60 psig balanced.
(Refer to the discussion in Section IV.C.3 below on balanced vs. unbalanced
seals.) Another disadvantage of outside mounted seals is that they are subject to exposure to dust and other environmental contaminants.
The port shown on the gland in Figure 5.14 could be used to inject an external liquid to lubricate the dynamic seal faces, in the event that the sealed
liquid pressure is below atmospheric or if it contains abrasives.
3. Single, Balanced Seals
A balanced seal is shown in Figure 5.13, with the pumped liquid coming
from the left side and atmosphere being on the right side. The purpose of a
balanced seal is to seal against higher pressures and speeds than possible in
an unbalanced configuration. The balance is achieved by having a stepped
or shortened primary ring and sleeve. These reduce the hydraulic pressure
at the dynamic seal faces, allowing the seal to handle higher pressure and
reducing the power consumption of the seal. The seals shown in Figures 5.10
and 5.12 are unbalanced seals, whereas those shown in Figures 5.11, 5.13, and
5.14 are balanced.
The maximum pressure that can be handled by a mechanical seal in an
unbalanced configuration is a function of the seal size (outside diameter of
the sleeve), the pump speed, the seal material, and the liquid being pumped.
Seal manufacturers provide curves or tables that show the maximum pressure that can be handled with unbalanced seals based on these four variables. If the seal must handle higher pressures than this, then a balanced seal
must be used.
4. Dual Seals
Dual mechanical seals are used for extremely tough services, where it is
desirable to completely eliminate the possibility of leakage of pumped product into the environment or where the dynamic seal faces must be isolated
from the pumped liquid. Examples of liquids for which a dual seal should
be considered include extremely corrosive liquids, toxic liquids, and liquids
containing abrasive solids. With the increasing amount of attention being
given to controlling or eliminating the release of liquid and vapor from
150
Pump Characteristics and Applications
pump seals because of environmental considerations, dual seals are considered by many as the best seal choice for this application.
Figures 5.15 and 5.16 show dual mechanical seals, with the liquid being
pumped coming from the left side of the figure and atmosphere being on the
right side. Dual seals include two dynamic seals, either mounted in tandem
(the arrangement shown in Figure 5.15) or back-to-back (as shown in Figure
5.16). In either case, they have a fluid zone between the two seals. Note that
both of the seals in Figures 5.15 and 5.16 are balanced designs.
There are two broad classes of dual seals, pressurized and unpressurized dual
seals, depending on whether the fluid zone between the two seals is at a pressure higher than the sealed liquid (pressurized dual seal) or at a very low
nominal pressure (unpressurized dual seal).
Note that earlier dual seal designs were referred to as double seals or tandem seals. Early double seals were mounted back to back, with a pressurized
fluid zone between the two seals, while early tandem seals were mounted in
tandem with an unpressurized fluid zone between the two seals. Today, seal
manufacturers offer a broader variety of design options. Dual seals mounted
back-to-back may have either a pressurized or an unpressurized fluid zone
between the seals. The same options are available for dual seals mounted in
tandem. So, the terms double and tandem have lost their significance and, in
fact, only cause more confusion in the already confusing seal terminology,
since the most important consideration is whether the fluid zone between
the two seals in a dual seal is pressurized or unpressurized.
Circulation of a suitable fluid at a
pressure greater (pressurized dual seal),
or lesser (unpressurized dual seal)
than the liquid sealed
FIGURE 5.15
Dual seal, with dynamic seals mounted in tandem. (Courtesy of John Crane Inc., Morton
Grove, IL.)
Sealing Systems and Sealless Pumps
151
FIGURE 5.16
Dual seal, with dynamic seals mounted back-to-back. (Courtesy of John Crane Inc., Morton
Grove, IL.)
If the fluid zone between the two seals is pressurized, it is referred to as
a barrier fluid, whereas if it is unpressurized, it is referred to as a buffer fluid.
A barrier fluid (liquid) is injected into the cavity between the two seals
of a pressurized dual seal, at a pressure higher than the pressure of the
pumped liquid at the seal. If the outside dynamic seal were to leak, the barrier liquid, rather than the pumped liquid, would leak out to the environment. If the inside dynamic seal were to leak, the barrier fluid would leak
into the pumped liquid, rather than the other way around. Thus, the two
dynamic seal faces are always kept completely free of the pumped liquid,
and the pumped liquid cannot possibly leak into the environment, provided that the barrier fluid remains pressurized.
The barrier fluid for a pressurized dual seal is chosen to be compatible with
the application, so that it may leak to the environment or into the pumped
product without severe negative consequences if either dynamic seal fails.
The barrier fluid is maintained at a pressure higher than pump pressure
at the seal by an auxiliary pump, or by the use of a vessel with pressurized
gas separated from the barrier fluid by a diaphragm. The loss of pressure on
the gas side of this diaphragm or a drop in the liquid level in the barrier fluid
vessel indicates a leak in one of the dynamic seal faces.
152
Pump Characteristics and Applications
An unpressurized dual seal has a different design consideration than a
pressurized dual seal. The unpressurized dual seal is meant to keep the
pumped liquid from leaking out to the environment, but it does this by providing a back-up seal, in the event of leakage of the primary seal. An unpressurized dual seal requires a buffer fluid (liquid), but the buffer fluid is not at
a pressure higher than the pumped liquid pressure, as is the case with the
pressurized dual seal.
Ordinarily, no pumped liquid reaches the secondary seal in an unpressurized dual seal, because the primary dynamic seal (along with its static
seals) keeps the liquid at the primary seal. A buffer fluid must be introduced
into the gland at the secondary seal to keep the dynamic seal faces of the
secondary seal lubricated. This buffer fluid can be at a low pressure, and is
sometimes simply supplied by a head tank mounted on the pump.
In the event that the primary seal fails, the secondary seal then takes over
to keep the pumped liquid from escaping to the environment. Usually, the
buffer fluid is monitored for pressure, level, pH, conductivity, or some other
convenient variable, to signify a failure of the primary seal.
The pump would not have to be shut down immediately if the primary seal
were to fail, however, because the secondary seal would then be sealing the
pumped liquid. The idea of the unpressurized dual seal is that if the primary
seal fails, the operator of the pump can plan an orderly shutdown of the pump
to repair the primary seal, without any product escaping to the environment.
The ability of an unpressurized dual seal to perform its intended service
of eliminating leakage to the environment depends on the promptness of the
operator in repairing the primary seal if it fails. If the pump is shut down for
a seal repair in a timely manner, the unpressurized dual seal works to eliminate any product leakage. If, however, the pump operates for an extended
time after the primary seal has failed, it is, during that period, simply operating as a single seal. If the secondary seal were to begin leaking during that
time, the pumped product could leak to the environment.
5. Gas Lubricated Noncontacting Seals
Gas lubricated noncontacting seals provide a different approach to a pressurized dual seal for controlling pump emissions, without the possibility
of leaking barrier fluid into the pumped liquid or into the environment. In
addition, they provide a reduction in power consumed at the seal. The seals
illustrated in Figures 5.17 and 5.18 are typical gas-lubricated noncontacting
seal installations, with the liquid being sealed coming from the left side of
the figure. Barrier gas from an external source (nitrogen, purified air, or other
inert gas) is supplied between the two seals of this pressurized dual seal, at a
pressure of 25 to 30 psi higher than the process liquid. When the shaft begins
to rotate, pressure is built up within spiral grooves that have been machined
into the mating rings, separating the seal faces. This hydrodynamic effect
results in noncontacting faces at both dynamic seal face positions.
Sealing Systems and Sealless Pumps
FIGURE 5.17
Gas-lubricated, non-contacting seal. (Courtesy of John Crane Inc., Morton Grove, IL.)
FIGURE 5.18
Gas-lubricated, non-contacting seal. (Courtesy of John Crane Inc., Morton Grove, IL.)
153
154
Pump Characteristics and Applications
The amount of barrier gas consumed is quite small. Horsepower losses at
the faces are extremely small because frictional contact has been eliminated.
This results in energy savings, as well as providing a seal that can achieve
zero emissions both in operation and during standby. Applications for gaslubricated noncontacting seals for pumps include abrasive, flashing, flammable, and toxic liquids.
6. Seal Piping Plans
There are a number of possible piping plans for how the seal flush piping
can be run for single and dual seals. The seal piping plans are indicated by
seal plan numbers, with the ANSI seal plan designation differing from the
API seal plan designation by the addition of a 73 (referring to the ANSI B73.1
specification for chemical pumps discussed in Chapter 4, Section XIV.B).
Thus, an API Plan 02 would be an ANSI Plan 7302. For convenience, the API
plan designations are used in this book.
Figures 5.19 to 5.25 show the most common seal piping plans. Plan 2 (Figure
5.19) is a noncirculating flush plan. Plan 11 (Figure 5.20) is the most common
seal flush plan in use today. This plan takes fluid from the pump discharge
(or from an intermediate stage) through an orifice, and directs it into the seal
chamber to provide cooling and lubrication of the seal faces. In a Plan 13
(Figure 5.21), the flow exits the seal chamber and is routed back to pump
suction. This is the standard arrangement for vertical turbine and high head
pumps. With a Plan 13, it is possible to control seal chamber pressure with
proper sizing of the orifice and throat bushing clearance. Plan 21 (Figure
5.22) is a cooled version of Plan 11 (Figure 5.20). The product from the pump
discharge is directed through an orifice, then to a heat exchanger to lower
the temperature before being introduced into the seal chamber. Note that
a Plan 22, not shown, is a modified Plan 21, with the addition of a strainer
before the orifice. Plan 32 (Figure 5.23) uses a flush stream brought in from
an external source to the seal, if the pumped liquid contains abrasives or
is otherwise not good for seal flush. Plan 52 (Figure 5.24) uses an external
reservoir to provide buffer fluid for the outer seal of an unpressurized dual
seal arrangement. Flow in the buffer fluid is typically induced by a pumping
ring. Plan 53A (Figure 5.25) uses an external reservoir to provide barrier fluid
for a pressurized dual seal arrangement. Flow in the barrier fluid is typically
induced by a pumping ring.
The key difference between Figures 5.24 and 5.25 is that the reservoir in
the unpressurized dual seal (Figure 5.24) is vented, while the reservoir in the
pressurized dual seal (Figure 5.25) is pressurized from a pressure source,
ussually nitrogen. Both figures show typical reservoir accessories (pressure
and level transmitters, and cooling coils).
Sealing Systems and Sealless Pumps
FIGURE 5.19
Seal piping Plan 2. (Courtesy of John Crane, Inc., Morton Grove, IL.)
FIGURE 5.20
Seal piping Plan 11. (Courtesy of John Crane, Inc., Morton Grove, IL.)
FIGURE 5.21
Seal piping Plan 13. (Courtesy of John Crane, Inc., Morton Grove, IL.)
155
156
Pump Characteristics and Applications
FIGURE 5.22
Seal piping Plan 21. (Courtesy of John Crane, Inc., Morton Grove, IL.)
FIGURE 5.23
Seal piping Plan 32. (Courtesy of John Crane, Inc., Morton Grove, IL.)
157
Sealing Systems and Sealless Pumps
Vent
FIGURE 5.24
Seal piping Plan 52 for unpressurized dual seal arrangement. (Courtesy of John Crane, Inc.,
Morton Grove, IL.)
Pressure source
FIGURE 5.25
Seal piping Plan 53A for pressurized dual seal arrangement. (Courtesy of John Crane, Inc.,
Morton Grove, IL.)
158
Pump Characteristics and Applications
V. Sealless Pumps
A. General
There are some pump applications where it is desirable or necessary to have
zero leakage of product, even less than the minimal leakage possible from
a well-chosen and well-maintained mechanical seal. The applications that
come to mind where even limited amounts of leakage would not usually be
tolerated include
• Toxic liquids (e.g., phosgene, cyanide)
• Hot liquids (e.g., heat transfer liquids)
• Carcinogens or environmental hazards
• Noxious, malodorous liquids
• Highly corrosive liquids
• Radioactive liquids
With the Clean Air Act and other federal and state environmental regulations beginning to be enforced more vigorously, the list of liquids that must
be pumped with zero leakage continues to grow.
Another incentive for eliminating shaft sealing completely is that packing
and mechanical seal servicing is probably the most frequent type of regular
maintenance that must be done on most pumps and thus is the cause of a
high percentage of overall pump maintenance expense. Included in the cost
of dual seal alternatives are the barrier or buffer fluids and the auxiliary
systems necessary to maintain them. Plus, some processes cannot tolerate
the possibility of contamination by barrier fluid leaking into the pumped
product from a pressurized dual seal.
Previous chapters in this book described several pump types that by their
design do not require a mechanical seal or packing assembly. Chapter 1 discussed two types of positive displacement pumps that are sealless: the peristaltic pump (Section VI.C.4) and the diaphragm pump (Section VI.C.13). Both
pump types, however, have limitations and shortcomings, chief among them
being their hydraulic coverage and the fact that they produce large pressure
pulsations. Still, if an application that requires sealless pumping falls within
the hydraulic range these pumps can handle and if the other characteristics
of these pump types that Chapter 1 described can be tolerated, they may be
a simple and relatively inexpensive sealless pumping choice.
Chapter 4, Section VIII discussed the vertical column pump, with one of
its advantages being that it, too, is a rather simple sealless pump. The pump
does have a shaft penetration at the top; but because the shaft column is normally relieved back to the sump in which the pump is immersed, the shaft
column is not pressurized. This pump type also has limitations, including
Sealing Systems and Sealless Pumps
159
hydraulic limitations, sleeve bearings often lubricated by the pumped liquid,
and the fact that to be sealless, the pump must be mounted over a tank or
sump, a configuration not always possible. Still, the vertical column pump
may represent a reasonable alternative for sealless pumping.
The following two subsections describe magnetic drive pumps and canned
motor pumps, two major types of sealless pumps that are generating a great
deal of interest and consideration, on the part of both pump manufacturers
and pump users.
B. Magnetic Drive Pumps
Magnetic drive pumps were invented more than 50 years ago and have
been available commercially for more than 30 years; but until the late 1980s,
their availability was largely limited to small nonmetallic pumps. There
were several European pump suppliers that offered heavier-duty metal process pumps earlier. By the late 1980s and early 1990s, driven largely by the
increased regulatory environment related to emissions from pumps, virtually all U.S. suppliers of centrifugal process pumps and a number of makers of positive displacement process pumps had introduced magnetic drive
technology.
Figure 5.26 illustrates a magnetic drive (hereafter called “mag drive”) process pump. The principle behind mag drive technology is that there is no
shaft penetration of the pressurized part of the pump casing. Instead, the
pump rotor is driven by magnetic flux. In Figure 5.26, the coupling end of the
pump is attached to an outer cylinder that includes a magnetic cylindrical
ring. The outer magnet is supported by the pump bearings shown between
the coupling shaft and the outer magnet and rotates with the motor. Smaller
versions are close-coupled, eliminating the ball or roller bearing system
shown in Figure 5.26.
There is a second magnetic cylinder located inside the pressurized part
of the pump casing and secured to the pump rotor. The outer and inner
magnetic cylinders are separated from each other by a containment can or
shell. The can is constructed of a corrosion-resistant alloy such as hastelloy,
or plastic in lighter-duty versions, and has a thin wall to minimize loss of
magnetic flux. The magnetic cylinder attached to the rotor is able to rotate
due to the magnetic flux between the inner and outer magnets. Thus, the
pump impeller rotates and liquid is pumped, without a shaft penetration
from the pressurized pump casing.
For all practical purposes, magnetic drive couplings exhibit an “on” or
“off” behavior; that is, both coupling halves are either running at the same
speed or are disconnected (decoupled). The halves are magnetically tied to
each other through the sealed can until such time that the torque capability
is exceeded. When the torque capability is exceeded, the pump stops quickly,
while the motor and driving magnet continue to rotate at full speed. The
halves will reconnect when the motor is stopped or nearly at zero speed.
160
Pump Characteristics and Applications
FIGURE 5.26
Magnetic drive process pump. (Courtesy of Goulds Pumps, Inc., a subsidiary of ITT
Corporation.)
Generally, manufacturers rate magnet decouple as the maximum torque
that the coupling will sustain as torque is raised gradually to the decouple
point. Sudden pressure pulses or action from valves in the system may result
in unexpected magnet decouple. High starting torque motors can exceed
the decouple point upon startup even though the pump has not reached
rated pressure. These factors must be taken into consideration when sizing
a magnet set. Also, the magnet decouple rating decreases with increasing
temperature.
Generally, it is not wise to depend on decouple to limit maximum pressure
of the pump, for example, in place of a safety relief valve for a positive displacement magnetic drive pump. Decouple ratings are generally published
as a minimum, and they can have a high tolerance range from magnet to
magnet.
Sealing Systems and Sealless Pumps
161
Advancements in magnet technology have helped move mag drive pumps
into larger and larger pump applications, with current upper limits being in
the range of 400 KW (535 HP). The most common materials used for the magnets in mag drive pumps are ceramic, samarium cobalt, and neodymium
iron boron. Each of these materials has different magnetic properties, cost,
availability, and upper temperature limits.
Ceramic magnets (barium or strontium ferrite) are the most common magnet
materials for smaller magnetic drive pumps. They are low cost, have excellent chemical resistance, moderate magnetic strength density, and high
sustained temperature capability (around 572°F, or 300°C). Their downside is a considerable torque loss with increasing temperature (0.2% to
0.5% per °C). They lose magnetism at a high rate as the temperature rises,
but they tend to sustain higher temperatures than some of the rare earth
magnets without permanent loss of magnetic strength. They are generally
considered low-temperature magnets, but one would be wise to run the
numbers before opting for the more expensive alternatives. If pressure
requirements are low, a ceramic magnet can be a very good choice for high
temperature.
Neodymium iron boron is a rare earth magnet material that is moderate in
cost, especially in high production. These magnets have very poor chemical
resistance (must be protected), very high magnetic strength density, and low
torque loss with increasing temperature, until their upper temperature limit,
which is about 250°F (121°C).
Samarium cobalt, another rare earth magnet, is relatively high in cost, has
good chemical resistance, very high magnetic strength density, low torque
loss with increasing temperature, and superior sustained high-temperature
capability (above 482°F, or 250°C).
It is important to understand that mag drives are not a panacea. Despite
the fact that they are taking the place of the high-maintenance mechanical
seal, there are a number of technological limitations that currently exist with
magnetic drives. The chief of these include the following.
1. Bearings in the Pumped Liquid
The pump impeller is still generating radial and thrust loads, just like any
centrifugal pump. Unfortunately, these loads cannot be passed through
the magnetic flux to bearings outside the pressurized pump casing. Mag
drive pumps have bushing-type radial bearings and plate or washer-type
thrust bearings (shown in green inside the can in Figure 5.26) that are usually exposed to the liquid being pumped. The pump must have a circulation path to allow pumped liquid to lubricate these bearings. Depending on
the material of the bearings and on the nature of the liquid being pumped,
the bearings are more or less subjected to abrasive wear and corrosion
from the pumped product. The bearings are usually made of a hardened
material such as silicon carbide or chrome oxide to resist abrasive wear and
162
Pump Characteristics and Applications
corrosion. Many mag drive pumps offer a screen built into the pump to
attempt to remove some of the abrasives before the liquid from the pump is
allowed to be introduced to the bearings. Most (but not all) manufacturers
make their impellers for mag drive pumps in closed configuration, which, as
discussed in Chapter 4, Section II.D, reduces axial thrust, compared with an
open impeller. Other efforts are made to balance out radial and thrust loads
as much as possible. Features such as external flush arrangements are available from some manufacturers.
Despite these design considerations, the bushing and plate-type bearings used on a mag drive pump are no match for the ball and roller type
bearings used to accommodate the radial and thrust loads of most centrifugal pumps. Also, the type of bearings used and the flow paths necessary
to provide lubrication for these bearings make a mag drive pump more
complex and include more wetted parts than most conventional centrifugal
pumps.
2. Dry Running
In addition to providing lubrication to the pump bearings, the flush liquid
circulating through the mag drive pump has a secondary function. A considerable amount of heat is generated due to the magnetic flux passing across
the metal containment can (especially for larger sizes), and this heat must be
carried away by the liquid circulating through the pump. Most mag drive
pumps do not tolerate running dry due to the heat given off from the magnetic flux, and a rapid failure occurs if flow into the pump is interrupted.
Bearing temperature detectors or other instruments are sometimes incorporated into the pump to help indicate a problem in this regard and shut down
the pump if a loss of flow to the pump is indicated, before the pump fails.
3. Inefficiency
Because of the loss of magnetic flux as it passes across the usually metallic containment shell, mag drive pumps are less energy efficient than their
conventional sealed counterparts. This means that if a sealed pump in the
field is being replaced with a mag drive pump, it is possible that a larger
motor might be required, which may in turn require a larger base plate. An
alternative design to significantly reduce these additional losses uses a plastic containment shell to avoid the flux losses that occur with the metallic
containment shell. This design alternative has had only limited acceptance,
however, due to misgivings on the part of industrial users about relying on a
nonmetallic shell for pressure containment.
4. Temperature
As mentioned previously, magnet materials have upper temperature limits.
Sealing Systems and Sealless Pumps
163
5. Viscosity
Because of the relatively small passageways for the recirculation of liquid
through the pump to lubricate the bearings, most mag drive pumps have
upper limits of viscosity that can be handled by the pump and still allow
acceptable amounts of bearing flush circulation.
Despite the shortcomings discussed above, mag drive technology offers
one solution to sealless pumping in process applications. The fact that so
many makers of process pumps, including virtually all ANSI and many API
pump makers, have committed so much effort and resources to developing
mag drive offerings is a sign of the commitment of the pump industry to this
technology. As magnet materials continue to improve, and as better ways
to combat the limitations discussed above are developed, mag drive pumps
will play a growing role in the process pump industry.
C. Canned Motor Pumps
Canned motor pump technology has been commercially available for more
than 50 years but is receiving a great deal of attention today as an alternative
to mag drive technology for process pump applications in which sealless
FIGURE 5.27
Canned motor pump. (Courtesy of Teikoku USA/Chempump, Warminster, PA.)
164
Pump Characteristics and Applications
pumping is required. Figures 5.27 and 5.28 illustrate a canned motor pump.
This pump technology relies on a specially designed motor, close-coupled
to the pump, and sealed at all connections to prevent any leakage. The rotor
portion of the motor is exposed to the pumped liquid for the same purpose
as the circulating liquid in a mag drive pump, namely to lubricate the radial
and thrust bearings required to support the pump. Figure 5.28 shows a
typical recirculation path to lubricate the bearing surfaces, though there are
alternative arrangements. The stator of the motor is separated from the rotor
by a metallic containment shell so that the motor stator remains dry.
Canned motor pumps have several limitations in common with mag drive
pumps. The most important one is that the radial and thrust bearings of the
canned motor pump must be wetted by the pumped liquid. Therefore, these
bearings must be sleeve and plate type, hardened to keep from prematurely
wearing from abrasive liquid. Manufacturers of canned motor pumps offer a
variety of designs for circulation paths through the pump and motor to provide cooling and lubrication of these bearings. There are also upper limits of
viscosity that are similar to those of mag drive pumps.
FIGURE 5.28
Canned motor pump lubrication flow path. (Courtesy of Teikoku USA/Chempump,
Warminster, PA.)
Sealing Systems and Sealless Pumps
165
The proponents of canned motor pumps offer the following advantages
compared with mag drive pumps.
1. Fewer Bearings
The mag drive pump must have a set of bearings in the pump motor, plus
(except in the smallest sizes) a set of bearings supporting the rotating magnetic cylinder. The canned motor pump does not require these extra sets of
bearings. Although this argument is true, it must be pointed out that the
extra bearings required by the mag drive pump are lightly loaded ball or
roller bearings and are not exposed to process liquids. Both pump types
have a radial and thrust bearing system that is exposed to the process
liquid.
2. More Compact
The close-coupled configuration of the canned motor pump takes up less
space than the bedplate-mounted mag drive pump and motor. The proponents of mag drive technology would argue that this is a less structurally
stable alternative than a bedplate mounted pump because the bedplate on a
mag drive pump can carry all loads down to the foundation.
3. Double Containment
The outer containment shell of a canned motor pump is constructed of pipe,
creating, along with the inner containment shell between the motor rotor
and stator, a double containment of the process liquid. Some, but not all,
makers of mag drive pumps offer double containment. However, the outer
containment shell is normally a cast part, with lower pressure-containing
capability than the canned motor’s outer containment pipe.
4. Lower First Cost
Based on their simplicity of construction, canned motor pumps should
be less costly to manufacture than mag drive pumps in equal quantities.
However, based on the commonality of some mag drive pump components
with higher volume (such as ANSI pump) components, mag drive pumps
may be less costly under some circumstances.
Offsetting the advantages of canned motor pumps is one fairly large shortcoming. If a canned motor pump has a problem, the user must return to the
manufacturer from whom the canned motor pump was purchased to get
repairs or replacement components because the motor is uniquely designed
and manufactured. In contrast, the motor for a mag drive pump is a standard
motor, available locally to the user on very short notice from a number of
manufacturers and suppliers. Thus, repair cost and time can be expected to
166
Pump Characteristics and Applications
be higher for a canned motor pump than for a mag drive pump. In areas of
high pump density, factory-authorized repair facilities have been established
by some manufacturers of canned motor pumps to address this shortcoming.
The upper limit of horsepower for canned motor pumps is currently
around 550 KW (737 HP), vs. 400 KW (535 HP) for mag drive pumps.
6
Energy Conservation and Life-Cycle Costs
I. Overview
This chapter describes a number of methods for reducing the amount of
energy consumed by centrifugal pumps. This includes methods of selecting
the right pump in the first place so that efficiency can be optimized, as well
as techniques for operating the pump to reduce energy consumption.
Maximizing pump efficiency by selecting the most efficient pump type
and operating the pump with minimal energy consumption are important
goals, particularly with larger horsepower pumps. For many pumps, small
and large, the cost of the energy to operate them over their lifetimes is much
more than their first cost or cumulative maintenance expense. The savings in
energy that can be achieved by selecting the most efficient pump for a given
service and operating it in the most efficient manner can represent a significant portion of total lifetime costs.
The importance of saving energy seems to ebb and flow in the minds of
pump users, depending on the cost and availability of energy supplies. Many
people have forgotten or were not around during the oil crisis of the 1970s,
when the cost of energy nearly quadrupled in a short time period. With inflation adjusted energy prices today being much lower than they were at their
peak some years ago, some people may become complacent about energy,
believing that relatively inexpensive supplies will always be available. This
will most certainly not be the case.
This chapter includes a detailed discussion of variable-speed pumping, a
technique being used more and more widely to reduce the energy consumed
in pumping, and variable-speed drives, devices most commonly used to
achieve variable-speed pumping.
In addition to lifetime energy costs, this chapter describes an analytical
approach to the other life-cycle costs (LCCs) associated with pump ownership.
167
168
Pump Characteristics and Applications
II. Choosing the Most Efficient Pump
It is possible to affect the efficiency, and thus the amount of energy consumed for a given flow and head, by the type of pump chosen for a particular application.
Figure 2.42 and its explanation in Chapter 2, Section XIII.C, indicated that it is
possible to maximize the BEP of a pump chosen for a particular application by
varying the number of stages or the speed of the pump chosen for the application (thus changing specific speed). However, choosing the most efficient pump
does not necessarily make it the lowest first-cost choice nor the choice with the
lowest operating costs. There may be other considerations too, such as space
constraints, ease of control, and noise. The following two examples illustrate
the use of Figure 2.42 in examining how varying the number of stages or speed
of the pump chosen can affect the efficiency. The examples show the type of
analysis that might be performed to select the pump that minimizes total cost.
Example 6.1: Selection of the Number of
Stages to Minimize Total Cost
PROBLEM
A pump must be chosen to meet a particular rating. Several alternative
pumps are analyzed, each with a different number of stages, to make the
rating. Then, the effect of the chosen pumps on energy, first cost, cost of
maintenance, space, and ease of control are examined.
GIVEN
Capacity = 500 gpm
Total head = 600 ft
Liquid = water at 60°F (SG = 1.0)
Pump speed = 3550 rpm
SOLUTION
Alternatives of single-, two-, and three-stage construction are considered. For each pump considered, the specific speed, Ns, is calculated,
using Equation 2.23. For the single-stage pump, the value of H (head per
stage) in the specific speed formula is 600; for the two-stage pump, it is
300; and for the three-stage pump, it is 200. Once specific speed is calculated, the expected efficiency at the BEP of the three alternatives can be
found using Figure 2.42. Then, using the efficiency for each pump, the
BHP for each pump can be computed, using Equation 2.16, with specific
gravity of 1.0. Table 6.1 shows the results of these computations:
Note that the table shows three quite different pumps to make the
same design rating point, all at the same speed of 3550, but with a varying number of stages.
169
Energy Conservation and Life-Cycle Costs
TABLE 6.1
Expected Performance: One-, Two-, and
Three-Stage Pumps
No. of Stages
Ns
Maximum
Efficiency (%)
BHP
1
2
3
655
1100
1495
68
77
80
111
98
95
Source: Figure 2.42.
The three-stage pump, being the most efficient, uses the least amount
of energy to operate. Per Table 6.1, the three-stage pump requires 3 HP
less than the two-stage pump and 16 HP less than the single-stage
pump. Using the technique described in Chapter 2, Section XII, with an
assumed power cost of $0.10 per kW-h and an assumed operating duty
cycle of 6000 h per year, the 16-HP difference between the three- and
single-stage pumps is computed to be worth about $7160 per year.
Note that a four-stage pump has not been considered in this problem,
because the amount of improved efficiency compared with a three-stage
pump is negligible. That is because the specific speed for the three-stage
pump places it near the flat part of the curve in Figure 2.42. Further
increases in specific speed would not greatly increase the expected
pump efficiency.
The problem statement suggests that other factors in addition to
energy should be evaluated. These factors, and the likely best pump
choice for each factor, are discussed below. It is clear from the discussion above and to follow that there is no best choice to optimize all the
factors. Hopefully, however, this example provides some guidelines for
making good engineering decisions as to the best choice for a given
pump application.
• First cost—The single-stage pump would likely be the lowest
first-cost alternative. Its configuration would be end suction,
whereas the two- and three-stage alternatives would be radially or axially split multistage pumps, which would likely cost
more. One factor that might lower the three-stage pump’s first
cost compared with the other two is the fact, per Table 6.1, that
with a lower required BHP, it might be possible that the threestage pump could use a 100-HP motor. (This assumes that the
pump will not operate at any higher capacity than the design
capacity, or that the motor service factor is used if the pump
operates at higher capacities.) The single-stage pump, on the
other hand, requires 111 HP to operate at the design capacity,
so it would possibly need a larger motor than the three-stage
pump. Still, this cost disadvantage of the single-stage pump is
not likely to change its status of being the lowest first cost of the
three alternatives.
170
Pump Characteristics and Applications
• Maintenance expense—Experienced maintenance people would
generally agree that the fewer the number of stages, the less
the expected maintenance expense for a pump. Fewer impellers
mean that it takes a shorter time to disassemble and reassemble
the pump, that there are fewer required spare rotating parts
(impellers, wear rings), the shaft is simpler, and that there are
fewer impellers to remachine or rebalance in the event work is
done on them.
An offsetting argument in favor of lower maintenance costs
for the three-stage pump is the fact that because the three-stage
pump makes the head with three impellers instead of one, each
of the three impellers is likely smaller in diameter than the singlestage impeller. This means that the pump owner might be able
to work on the smaller impellers in an in-house machine shop
(e.g., trimming impellers, making wear rings, balancing impellers, etc.). The single-stage impeller, on the other hand, might
have to be sent to an outside shop for the same operations,
which could be more expensive. This argument, while having
some merit, is not likely to offset the statement in the previous
paragraph that the pump with the fewest number of stages is
likely to have the lowest maintenance expense.
• Space—Space allowed for the pump may be an issue for a plant
modification, for a skid-mounted assembly, in tight or confined
spaces such as on ships or offshore platforms, and for machines
or systems that include a pump as a part of the operation. If
space is an issue, either the single-stage or the three-stage
pump might be the best choice, depending on whether the
space limitation is on the diameter or on the length. The singlestage pump would have the largest diameter but the shortest
length, while the three-stage pump would be just the opposite,
with the smallest diameter but the longest length.
• Ease of control—The material in Chapter 2, Section VII, on specific speed indicates that the higher the pump specific speed,
the steeper the resulting pump H–Q curve. And the discussion
in Chapter 2, Section IX, on system head curves mentions the
fact that a steeper H–Q curve is easier to control because it does
not produce wide flow swings with variations of system head.
Therefore, if a more constant flow is required without resorting
to a flow control valve (FCV), then the higher specific speed
pump (the three-stage pump) would be the best alternative.
Example 6.2: Selection of Pump Speed to Minimize Total Cost
PROBLEM
A pump must be chosen to meet a particular rating. Two alternative singlestage pumps are analyzed, one at 1780 rpm and one at 3550 rpm, to make the
rating. The likely effect of the chosen pumps on energy cost, first cost, maintenance expense, space, ease of control, and noise level are then discussed.
171
Energy Conservation and Life-Cycle Costs
GIVEN
Capacity = 500 gpm
Total head = 175 ft
Liquid = water at 60°F (SG = 1.0)
Number of stages = 1
SOLUTION
For each pump considered, the specific speed, Ns, is calculated using
Equation 2.23. Once the specific speed is calculated, the expected BEP
efficiencies of the two alternatives can be found using Figure 2.42.
Finally, using the efficiency for each pump, the BHP for each pump can
be computed, using Equation 2.16, with specific gravity of 1.0. Table 6.2
lists the results of these computations.
The two alternative pumps considered in this example each make the
desired rating point with one stage, but one of them does it while operating at 1780 rpm, the other while operating at 3550 rpm. They are two
completely different pumps, each making the same rating point at a different running speed.
The two-pole (3550 rpm) pump results in the highest efficiency and the
lowest energy cost. Does this mean that two-pole pumps should be used
for all applications? Not necessarily, because per the problem statement,
other factors in addition to energy costs must be considered. They are
discussed below. Noise level concerns favor slower-speed pumps, but
noise is often not a concern for industrial pumps, especially in relatively
small sizes such as this example. The following discussion shows that
the only significant factor in favor of the slower-speed pump is maintenance expense. However, this factor is an extremely important one, and
may in fact be the deciding one. If the pump owner has experience with
two-pole pumps in this size, however, and is comfortable with the alignment and balance levels necessary for two-pole pumps, then the higherspeed pump is likely to be the choice representing the lowest total pump
cost after factoring in all the variables.
• First cost—The two-pole pump would be the lowest first-cost
machine because the pump required to make the rating at twopole speed would be quite a bit smaller than the pump required
to make the rating at four-pole speed. (This assumes that an end
suction pump would be used in either case.) As an example, if
TABLE 6.2
Expected Performance: 1780- and
3550-rpm Pumps
Speed
Ns
Maximum
Efficiency (%)
BHP
1780
3550
825
1650
72
80
31
28
Source: Figure 2.42.
172
Pump Characteristics and Applications
ANSI pumps were used for this application, the two-pole pump
selected would be a 3 × 4 – 7, whereas the four-pole selection
would be a 3 × 4 – 13 (see Figure 2.8).
The two-pole pump motor also would be less expensive
because it too would have a physically smaller frame than the
four-pole motor. For a given motor horsepower size, the higher
the speed, the smaller the required frame size.
• Maintenance expense—In general, the higher-speed pump
should cost more to maintain. There are several reasons for
this. At higher speeds, pumps are less tolerant of misalignment
and imbalance than they are at lower speeds. Also, at higher
pump speeds, the seal’s dynamic faces would see higher surface speeds and twice as many rotations per year, so the seal
life on the higher-speed pump would likely be shorter. This is
somewhat offset by the fact that the higher-speed pump would
have a lower torque than the same horsepower pump at a lower
speed, as per Equation 2.18. Therefore, the two-pole pump
could get by with a smaller shaft diameter, and thus a smaller
seal size. Another offsetting argument against the high-speed
pump having the higher maintenance cost is the fact that the
components of the high-speed pump are physically smaller
than the same parts for the slower-speed pump. A replacement
impeller or wear ring for the four-pole pump would cost more
than the comparable part for the higher-speed pump.
Despite arguments to the contrary, most people experienced
with pump operation and maintenance would agree that wear
and tear on a two-pole pump is higher than that on a four-pole
pump. Many users simply do not use two-pole pumps because
of concerns about maintenance. On the other hand, users at
other plants such as refineries and power plants are quite comfortable with pumps operating at this speed, and have good
pump alignment and balancing techniques, so a pump at twopole speed would not be a concern for them at all.
Note that the consideration of two- or four-pole speed
assumes that the liquid being pumped is relatively clean.
Applications involving paper stock or abrasive slurries are generally chosen at speeds slower than 1800 rpm in an effort to
reduce excessive erosion in the high-velocity areas.
• Space—As covered in the discussion of first cost, the twopole pump and motor would be smaller than the four-pole
alternative.
• Ease of control—With the higher specific speed, the two-pole
pump would have a steeper H–Q curve and thus would be easier to limit flow swings with system head variations.
• Noise level—Most of the noise from a centrifugal pump is
caused by the motor. In general, for the same motor horsepower, the higher the speed, the higher the noise level. Thus,
the two-pole pump would have a higher noise level than the
four-pole pump. This may not be a problem for most industrial
pumps, especially in the relatively small size of this example. It
Energy Conservation and Life-Cycle Costs
173
could, however, be a concern if the pumps were to be operated
in an area near the general public or where noise levels need to
be otherwise minimized.
• NPSH—If NPSH is an issue, the slower-speed pump would be
the favored alternative, as it would have a lower NPSHr than
the higher-speed pump.
The preceding examples focused on the tools that allow an engineer or
pump user to select the most efficient pump for a given application, while
being mindful that a selection based only on efficiency can also affect initial
cost, maintenance, and operating parameters for the pump.
In the above examples, the pump head is a given value. It is the responsibility of the person who is sizing the pump and designing the system to determine the value of pump head. As demonstrated in Chapter 2, Section XII,
oversizing the pump can lead to a considerable waste of energy.
Another point to consider in choosing equipment for minimal energy consumption is today’s availability of premium-efficiency motors. These motors
are discussed in Chapter 4, Section XVI.
III. Operating with Minimal Energy
Once a particular pump has been chosen for an application, the best way
to minimize energy consumption in the operation of the pump is to keep
the pump operating as efficiently as possible. Make certain that the impeller
settings for open impellers and the wear ring clearances for closed impellers are at their minimum recommended amounts considering the impeller
size, materials of construction, any galling tendencies for the materials, temperature, and amount of abrasives present. (Recommended axial settings for
open impellers and wear ring clearances for closed impellers are discussed
in Chapter 4, Section II.A) Impeller clearances should be reset as often as
reasonable to achieve minimal leakage of liquid back to suction.
A centrifugal pump should not be operated at higher flow rates than
required for the process. Remember, with most centrifugal pumps, the
higher the flow rate, the higher the horsepower consumed.
Any throttling being done in the system (by valves or orifices) should be
examined carefully, because this is a source of wasted energy. It may be that
the pump has been oversized for the system requirements. An impeller trim
or some other change such as a speed reduction can eliminate the unnecessary throttling in the system. (Refer to the discussion on the effects of oversizing pumps in Chapter 2, Section XII.)
If a pump must operate over a wide range of flow and head at different
times, consideration should be given to multiple pumps operating in parallel
or in series, as discussed in Chapter 2, Sections X and XI. Another possible
174
Pump Characteristics and Applications
solution if a multistage pump is being used might be to de-stage the pump
during the time when head and flow requirements are lower. This is, for
instance, commonly done with vertical turbine pumps in the mining industry. A third, more flexible alternative, variable-speed operation, is discussed
in the following section.
IV. Variable-Speed Pumping Systems
Variable-speed pumping has been around for many years, but its use has
become more readily justified in recent years due to improvements in the
technology for achieving variable-speed control of pumps and the reduction in the cost of such devices. Also, efforts by electric utilities to help their
commercial and industrial customers reduce energy consumption through
demand-side management have included incentives to incorporate variable
speed into pump systems.
Variable speed is most easily justified when the pump must deliver a
wide range of flow over time. Examples of pumping systems that demand
a range of flow include process pumps in a variable-capacity process plant,
municipal water and wastewater pumps, HVAC chilled water, and cooling
water pumps in commercial and institutional buildings, pipe line pumps,
and power station pumps. In general, the wider the range of flow demand,
the greater the likelihood that variable speed can be justified. However, an
extremely wide range of flow is not necessarily required to justify the use of
a variable-speed pumping system, since a 10% decrease in flow can reduce
the power requirement by approximately 27%.
In addition to applications that require variable flow, there are also a number of applications that require constant flow, but can still benefit from controlling the speed of the pump. The most common of these are applications
that require a variable head pressure to deliver a constant flow. One example
of this is a swimming pool filtration loop circulation pump. Another is a
cooling tower filtration system that pumps water from the basin of a cooling
tower, through a strainer, and back to the tower basin. If a constant-speed
pump is used in either of these applications, the flow of filtered water will
decrease as the filter loads. An adjustable-speed drive can increase the speed
of the pump, making it possible to maintain a constant flow over a wide range
of filter conditions. An additional benefit of this system is that variable-speed
control can even indicate when the filter needs to be cleaned or replaced by
signaling when the pump’s speed is near maximum.
Primary hot or chilled water pumps and other pumps that normally
provide constant flow against a constant head can also be candidates for
variable-speed pumping. During the commissioning of the system, it is
often necessary to reduce the capacity of the pump to achieve the proper
175
Energy Conservation and Life-Cycle Costs
temperature differential and flow. Although this could be done by disassembling the pump, trimming its impeller, and reassembling the pump, this
is not usually done. Instead, it is common to simply manually close a throttling valve to reduce the flow to the desired level. Doing this can introduce a
significant, constant energy loss in the system. It can be economical to simply
leave the valve fully open and manually adjust the speed of the pump to
achieve the desired system performance. An additional advantage of using a
variable-speed drive is that the flow can be easily readjusted in the future if
the operating conditions of the system change. This would be more difficult
to accomplish if flow were adjusted by trimming the impeller of the pump.
Variable-speed pumping saves energy by directly controlling the capacity
of a pump by changing the pump’s speed, rather than running the pump
at full speed and using a valve to externally restrict or bypass excess flow.
Using a variable-speed drive is much like driving a car down a highway and
controlling its speed by adjusting the position of the gas pedal. By doing this,
the output from the engine is directly controlled to meet the requirements of
the system. In contrast, using a throttling valve to control flow is much like
driving a car down a highway by keeping the gas pedal floored and using
the brakes to control the car’s speed.
The energy savings that result from using a variable-speed drive are illustrated in Figures 6.1–6.6. When the speed of the pump is held constant, pump
curve A in Figure 6.1 shows how this pump operates between flows 1 and 2.
The higher flow rate is delivered by the pump when pump curve A intersects the system resistance curve X at point ax. The power associated with
this flow is proportional to the flow times the pressure. A convenient way to
visualize this is by observing the area of rectangle with a vertex at point ax.
X
H (pressure)
ax
A
Power
1
Q (flow)
2
FIGURE 6.1
Power consumed in a pump is proportional to flow times pressure. (Courtesy of Danfoss
Graham, Milwaukee, WI.)
176
Pump Characteristics and Applications
Y
ay
X
H (pressure)
ax
A
Power
1
2
Q (flow)
FIGURE 6.2
Using a throttling valve on a constant-speed pump effectively steepens the system resistance
curve. (Courtesy of Danfoss Graham, Milwaukee, WI.)
To achieve the lower flow rate 1, a valve at the discharge of the pump can
be closed. This increases the output pressure required by the pump and
decreases the flow, as shown by system resistance curve Y and point ay in
Figure 6.2. The power associated with this flow is illustrated by the area of
the rectangle with a vertex at point ay. Although this power is normally less
than the power required for full flow (except for the case of an axial flow
pump), there is a significant amount of waste. This wasted power is illustrated in Figure 6.3.
Y
ay
X
H (pressure)
ax
A
Extra power
required for system
resistance curve Y
1
Q (flow)
2
Power
required for system
resistance curve X
FIGURE 6.3
Illustration of power wasted by throttling. (Courtesy of Danfoss Graham, Milwaukee, WI.)
177
Energy Conservation and Life-Cycle Costs
The amount of power required to produce flow 1 on system resistance
curve X is illustrated by the area of the bottom rectangle in Figure 6.3. The
area of the top rectangle shows the additional power that is associated with
creating the same flow while following system resistance curve Y. The additional power is required to overcome the pressure drop across the throttling
valve that was added to the system.
Throttling the output of the pump can waste energy in two ways. First, if
the pump was operating near its BEP at point ax, then at point ay it is operating at a reduced efficiency. Second, the pump is required to produce an
increased head when it is producing the reduced flow. The system resistance
curve shows that the system requires reduced pressure for reduced flow. The
additional pressure drop is simply the pressure lost across the throttling
valve. The power required to produce this pressure drop at this flow represents wasted energy.
Controlling the speed of the pump both helps to maintain high pump efficiency and to reduce energy consumption. When the speed of the pump is
decreased, the flow is reduced by following the system resistance curve X to
the lower flow in Figure 6.4. Instead of being forced to follow the full speed
pump curve, adjusting the speed of the pump generates a new pump curve
for each new speed. The system will now operate at point bx in Figure 6.4.
Because no artificial pressure drops are imposed on the system to reduce
the flow, the energy loss is minimized. The area of the associated rectangle
shows that the power associated with this flow is significantly less than the
power that was required when a throttling valve was used to reduce flow.
Although controlling the speed of the pump will always result in the optimum amount of energy saving, the actual energy savings that will result
A
X
Pump curve A
for high speed
H (pressure)
ax
Pump curve B
for low speed
B
bx
1
Q (flow)
2
Power required
for low pump speed
FIGURE 6.4
Using a slower pump speed to obtain a lower flow rate. (Courtesy of Danfoss Graham,
Milwaukee, WI.)
178
Pump Characteristics and Applications
depends on the pressure requirements of the system, as indicated by the
system resistance curve.
If no pressure is needed at zero flow (see Figure 6.5), the energy savings
will be the greatest. In this case, the required pressure will proportional to
the flow squared.
H ∝ Q2
(6.1)
HP ∝ H × Q
(6.2)
HP ∝ Q3
(6.3)
As a result, because
Then,
H (pressure)
This is the centrifugal pump affinity law, discussed in Chapter 2, which
predicts that reducing the flow to 50% of maximum will reduce the power
required to 12.5% of maximum. The analysis above only applies to systems
whose pressure requirements approach zero when flow is reduced toward
zero as shown in Figure 6.5. In other situations, the system resistance curve
looks like the one in Figure 6.6. Clearly, the power required to produce a
given reduced flow with that system is greater than it would be for a system
with no minimum pressure requirement. While controlling the speed of the
pump will provide the greatest energy savings, the static head requirement
of the system will dictate the maximum energy savings potential.
A
Power for
100% flow
Power for
50% flow
0%
50%
Q (flow)
100%
FIGURE 6.5
Illustration of power requirement for 100% and 50% flow in a system with no static head.
(Courtesy of Danfoss Graham, Milwaukee, WI.)
179
H (pressure)
Energy Conservation and Life-Cycle Costs
A
Power for
100% flow
High
static head
Power for
50% flow
0%
50%
Q (flow)
100%
FIGURE 6.6
Illustration of power requirement for 100% and 50% flow in a system with some amount of
static head. (Courtesy of Danfoss Graham, Milwaukee, WI.)
Open pumping systems, such as lift station pumps and pressure booster
pumps for potable water, often have a significant static pressure head
requirement. Closed-loop pumping systems, such as hot water and chilled
water pumping loops in heating, ventilation, and air conditioning systems
often have a minimum pressure set point requirement imposed by the control system to ensure proper operation under all flow requirements. This
pressure set point acts like the static head in an open pumping system. To
maximize energy savings, it is important to use the lowest system set point
that provides acceptable system operation.
The reasons why the use of variable speed in a pumping system saves
energy are further illustrated in Figure 6.7. The pump designated as curve A
must operate between the flows marked as 1 and 2 in the figure. The higher
flow rate delivered by the pump occurs where the pump H–Q curve intersects the system resistance curve X at point ax. Using a constant-speed pump,
the lower flow rate is achieved by throttling the pump (i.e., closing a valve
at the pump discharge) to steepen the system resistance curve until the new
system resistance curve (curve Y) intersects the pump curve at the lower
flow rate, at point ay in Figure 6.7.
Two things happen at the lower flow rate using a constant-speed pump. First,
the pump is operating at a lower efficiency at this lower flow rate. (Refer to the
efficiency curve A in Figure 6.7.) Second, the pump head at point ay (flow rate
1) is higher than the head at point ax (flow rate 2) because the constant-speed
pump must operate on its characteristic head–capacity curve.
If instead of throttling, the pump is slowed down, it would produce new
head–capacity and efficiency curves, shown in Figure 6.7 as curves B, which
180
Pump Characteristics and Applications
A curves - High speed
B curves - Low speed
Efficiency
(%)
B
Y
A
X
ay
ax
H
(ft)
bx
A
B
1
Q (gpm)
2
FIGURE 6.7
Pump H–Q and efficiency curves at two speeds, and system curve before and after throttling.
would follow the affinity laws. (Refer to discussion of the affinity laws in
Chapter 2, Section VIII.) The intersection of this slower speed head–capacity
curve (labeled curve B in Figure 6.7) with the original system head curve X
(point bx in Figure 6.7), represents the flow rate achievable by operating the
pump at reduced speed without throttling. The difference in head between
points ay and bx represents the head wasted across the control valve when
the constant-speed pump is throttled instead of slowed down. Note that the
efficiency curve for the pump at the lower speed shifts to the left. The pump
still runs near its BEP when operating at the lower speed, provided the system curve is one similar to Figure 6.7. With a flatter system curve, the pump
moves further to the left of BEP.
The use of variable speed instead of throttling a constant-speed pump
saves energy in two ways in the above example. The pump operates near
its best efficiency point, and no energy is wasted across the throttling valve.
There are other benefits in addition to energy savings, as are detailed below.
Variable-speed pumping systems have evolved in their method of creating
speed changes. Several pump driver types, such as steam turbines, hydraulic
motors, air motors, and diesel engines, are inherently able to achieve a range
of pump speeds. Even simpler mechanical devices such as gear boxes and
belt drives can achieve incremental, although not variable, speed changes.
Magnetic slip has also been used to produce variable-speed pumping.
Initially, this was done by altering the motor’s slip by externally increasing
Energy Conservation and Life-Cycle Costs
181
the resistance of its rotor. Although this allowed the speed of the motor’s
shaft to be adjusted, the energy lost through the external resistors and the
maintenance of the brushes needed to connect these resistors to the motor’s
rotor were significant concerns. Eventually this system was replaced by
external magnetic slip devices such as eddy current drives. Because these
were coupled between the motor and the load, mechanical mounting was
required. This required additional mounting space and made it particularly
difficult to retrofit a constant-speed system for variable speed. In addition, the
absolute maximum theoretical efficiency of such a system is equal to the percentage of full output speed at which the pump is driven. So, when the pump
was driven at 80% speed, over 20% of the energy delivered to the motor’s
shaft was wasted. This is a concern when energy conservation is the goal.
The bearings required to support the shafts in the drive also require periodic
maintenance.
The most efficient method of continuously adjusting the flow from a pump
is to directly control the speed of the electric motor that drives the pump.
Initially, it was only possible to efficiently control the speed of DC motors.
This was relatively simple because the speed control only had to adjust the
DC voltage that was applied to the motor while monitoring motor current.
The drawbacks of such a system were the cost of the motor and the periodic
maintenance that its brushes required.
When electronic power technology advanced sufficiently to allow the control of the speed of AC induction motors, the combination of an inexpensive,
low-maintenance motor with a high-efficiency speed control produced an
ideal package for controlling the flow from a pump. As a result, AC adjustable
frequency drives or variable frequency drives (VFDs) have become the standard
method for controlling the capacity of a pump.
Here is a brief overview of how a variable frequency drive operates. An AC
induction motor’s rotor (Figure 6.8) is driven by a rotating magnetic field that
is produced by its stator coils. This is done by applying alternating current
to the stator coils. As the alternating current in the stator coils changes, the
induced magnetic field also changes. When three-phase electricity is applied
to the coils, the resulting magnetic field rotates smoothly around the motor.
The frequency of the alternating current that is applied to the stator coils
controls the rotational speed of this magnetic field, and thus the speed of the
rotor. AC induction motors that are connected to the AC power line are singlespeed devices because the frequency of the power mains is fixed. A variablefrequency drive allows the frequency of the power applied to the motor to be
controlled. This allows the speed of the motor’s shaft to be adjusted.
In addition to controlling the frequency of the power applied to the motor,
a VFD must also control the voltage at the motor’s terminals. This is because
of the inductive reactance of the motor’s coils. Inductive reactance, XL, measures the resistance that the coils offer to the flow of AC current. This varies
with the frequency of alternating current applied to the coil according to the
following equation:
182
Pump Characteristics and Applications
Stator coil
Rotating
magnetic
field
Rotor
Stator coil
FIGURE 6.8
An AC induction motor. (Courtesy of Danfoss Graham, Milwaukee, WI.)
XL = 2πf L
(6.4)
From this formula it is clear that the inductive reactance, XL, of a coil is
low when the applied frequency, f, is low. As a result, if nothing else was
changed, a motor would tend to draw more current as the frequency applied
to it was reduced. To keep this from happening, voltage applied to the motor
is controlled along with the frequency.
For full-speed operation, the drive applies rated frequency and voltage
to the motor. For reduced-speed operation, the drive reduces both the frequency and the voltage of AC applied to the motor (Figure 6.9).
Some VFDs use a simple direct proportion to determine the motor’s voltage. This is generally used when a constant torque is required at all operating
speeds. This is often the case for positive displacement pump applications
(Figure 6.10).
High speed
Low speed
FIGURE 6.9
In a VFD, the voltage applied to the motor is controlled along with the frequency. (Courtesy of
Danfoss Graham, Milwaukee, WI.)
183
Drive output voltage
Motor output torque
Energy Conservation and Life-Cycle Costs
Drive output frequency
Drive output frequency
FIGURE 6.10
Drive output voltage–frequency pattern for constant motor torque. (Courtesy of Danfoss
Graham, Milwaukee, WI.)
Drive output voltage
Motor output torque
When a centrifugal pump is used, the torque that the motor must produce
at slow speeds is generally quite low. Providing a constant ratio between
motor voltage and frequency will produce a greater magnetic field in the
motor than is needed to drive the pump at low speed. This extra magnetizing current produces extra heat in the motor and thus reduces system efficiency. To maintain high efficiency throughout the speed range, some drives
have a “variable torque” setting. This reduces the voltage applied to the
motor to a greater extent at low speeds, minimizing the unnecessary motor
current (Figure 6.11). Some advanced VFDs even offer an automatic energy
optimizer feature. This automatically matches the drive’s output voltage to
the speed and torque requirements of the load.
Although a number of methods are used to control the voltage and frequency produced by a VFD, there are some basic similarities among all the
common variable-frequency drive designs (see Figure 6.12).
Drive output frequency
Drive output frequency
FIGURE 6.11
Drive output voltage–frequency pattern for reduced low-speed motor torque. (Courtesy of
Danfoss Graham, Milwaukee, WI.)
184
Pump Characteristics and Applications
DC
AC
Rectifier
AC
Inverter
FIGURE 6.12
A modern adjustable frequency drive consists of three sections. (Courtesy of Danfoss Graham,
Milwaukee, WI.)
First, the rectifier converts the incoming alternating current (AC) into direct
current (DC). One purpose of this conversion step is to remove the line frequency from the power. In this way, the drive’s output stage does not have to
continually compensate for the variations in the incoming alternating current. The DC gives the output stage of the drive clean power to start from
when it generates the output frequency that is needed for the desired output
speed. It is interesting to note that the DC bus voltage that is produced is
much closer to the peak voltage of the AC power line than it is to the average
voltage that is generally used to describe the AC line voltage.
The middle section of a modern VFD is the DC bus. The main purpose of
this is to filter out most of the residual “ripple” from the rectified AC line
power, so that there is no significant interaction between the AC power line’s
frequency and the drive’s output frequency.
This filtered DC is then fed to the drive’s output section, which is called
the inverter. Transistorized switches in the inverter direct the DC bus voltage
to the appropriate motor lead. By alternately connecting each motor lead to
the positive side of the DC bus and then to the negative side in the proper
sequence, an appropriate AC output is applied to the motor’s leads. The faster
the polarity reversals take place, the higher the output frequency, and so the
faster the motor’s speed.
Electrically, a “low-voltage” drive is defined as one that operates on an AC
line voltage of 600 V rms or less. The vast majority of modern low-voltage
VFDs are voltage source, pulse-width modulation (PWM) drives. The diagram
in Figure 6.13 fills in some details for a basic PWM drive.
The rectifier section consists of a set of diodes. These simply act as “check
valves,” forcing the input current in the proper direction to produce DC. It is
important to note that the DC bus voltage is not controlled; its value fundamentally depends on the incoming AC line voltage.
The DC bus consists of a bank of capacitors. The larger the size of the drive,
the greater the number of capacitors. These drives are called “voltage source”
185
Energy Conservation and Life-Cycle Costs
Diodes
change
AC to DC
Capacitors
filter the
DC
Transistors
switch DC
to AC
FIGURE 6.13
A pulse width modulation (PWM) variable-frequency drive. (Courtesy of Danfoss Graham,
Milwaukee, WI.)
because the capacitors act to filter the voltage of the DC bus, and attempt to
maintain it at a constant level. This filtered DC bus is also used to provide
control voltage to the rest of the drive. Because the DC bus capacitors tend to
maintain their voltage throughout a short power loss, such drives have the
ability to “ride through” brief line voltage sags with no significant interruption in operation.
The inverter section of a PWM variable-frequency drive consists of a set of
switching transistors. Each motor lead is connected to at least two transistors.
One transistor connects the positive side of the DC bus to the motor lead. The
other transistor connects the negative side of the DC bus to the motor lead.
A complete inverter consists of at least six transistors, one pair for each lead
of the three-phase motor. By controlling the switching of the transistors, the
frequency of polarity reversal of each motor lead can be set. This determines
the frequency of the AC that is applied to the motor, and thus its speed.
In addition to controlling the polarity of the voltage that is applied to the
motor, the inverter section of a PWM drive also controls the average voltage
that is applied to the motor. This is done by sending pulses of voltage to
the motor. When only a small amount of voltage is required by the motor,
the pulse is turned on for a very short period of time. The average voltage
of this narrow pulse is quite low. When a higher voltage must be applied
to the motor, the pulse is turned on for a longer period of time. This wider
pulse has a higher average voltage. Because this drive controls the average
motor voltage by controlling the width of the pulses that are applied to the
motor, such drives are called pulse width modulation, or PWM, drives. The
frequency of the pulses applied to the drive is generally in the range of 2 to
20 kHz. This is called the drive’s switching frequency or carrier frequency. The
range of common switching frequencies is large because each carries its own
186
Pump Characteristics and Applications
advantages and disadvantages. The specific application generally dictates
the ideal switching frequency.
In addition to controlling the voltage applied to the motor based on the
output frequency of the drive, modern PWM drives also continually adjust
the width of the pulses that are applied to the motor to simulate the smooth
increase and decrease of voltage that a sine wave AC voltage would apply
to the drive. Such “sine-coded PWM drives” simulate the AC voltage that
a pure AC power source would apply to the drive, ensuring efficient motor
operation. Some manufacturers of variable-frequency drives have developed
their own proprietary voltage control algorithms to optimize motor performance and minimize motor heating.
It seems strange initially to apply pulses of voltage to the motor rather than
use a transistor to “throttle” the voltage applied to the drive, producing a
smooth, sine-wave output voltage. This is not done for two reasons.
First, it would be quite inefficient. Using a transistor to smoothly throttle
back voltage while providing a significant amount of motor current would
be at least as wasteful as using a throttling valve to control the flow in a
pumping system. The only difference would be that the inefficiencies would
be moved from the pumping system to the variable-frequency drive. Because
the main purpose of using a VFD to control a pumping system is to maximize system efficiency, this would be quite unacceptable. By contrast, the
energy loss associated with switching an inverter’s transistors on and off at
a high switching frequency is quite small.
Second, it is not necessary. The important concern for motor operation is
the current that flows through its windings because it is the current that
produces the magnetic field that creates torque in the motor. The inductance
of the motor’s stator windings filters the motor’s current, making it closely
resemble a sine wave.
Insulated-gate bipolar transistors (IGBTs) are commonly used in the inverter
section of modern PWM variable-frequency drives. These have become the
modern standard for low-voltage drives because they are very reliable, have
a high efficiency, and can produce the high output switching frequencies
that are required.
Medium-voltage systems are defined as those with an AC line voltage
greater than 600 V AC and less than 38 kV AC. It is difficult to obtain IGBTs
that function well with the high voltages and currents associated with such
applications. So, it is generally necessary to control these drives using thyristors, which are also known as silicon controlled rectifiers (SCRs). These SCRs
cannot be switched as quickly as IGBTs, so they are of little use in generating
a PWM waveform to control voltage to the motor. Instead, a set of SCRs is
used in the input rectifier to control the voltage and current that is applied to
the DC bus. A large series inductor is then generally used to smooth the current in the DC bus. Because this inductor acts to regulate the current passing
through it, these drives are called current-source drives. Finally, a set of SCRs
is used in the inverter section to switch the controlled DC bus to the motor.
187
Energy Conservation and Life-Cycle Costs
Such drives are more cumbersome than the PWM drives that are common
for low-voltage drives. However, they are used in medium-voltage applications because appropriate IGBTs are not readily available.
The following example, courtesy of Danfoss A/S, Milwaukee, WI, illustrates the energy savings that can be achieved using a VFD in the condenser
pump system shown in Figure 6.14 with a traditional throttling valve. For
this system, a load profile, showing the percent of maximum flow required
to satisfy the condenser loads during the various times of the day or days
of the year, is prepared (see Figure 6.15). As an alternative to throttling the
Water
inlet
Chiller
Basin
Water
outlet
Condenser
water
pump
Throttling
valve
FIGURE 6.14
Traditional condenser pump system with throttle valve. (Courtesy of Danfoss A/S, Milwaukee,
WI.)
40
35
% Operating hours
35
30
25
20
25
20
15
15
10
5
5
0
60
70
80
90
100
% Maximum volume flow rate
FIGURE 6.15
Load profile showing percent (%) operating hours and percent (%) of maximum flow rate.
(Courtesy of Danfoss A/S, Milwaukee, WI.)
188
Pump Characteristics and Applications
pump, a VFD controlled via a flowmeter is being considered (Figure 6.16).
The system head curve and full-speed pump curve are shown in Figure 6.17.
Three comparisons are presented for a 40-HP/30-kW condenser water pump
with the system and load profile just described. The energy consumption
during one year of operation is calculated for each. The comparisons are
shown is Figure 6.18.
In the first calculation, a 15% overheaded pump uses a discharge balancing
valve to adjust the pump flow to the required system design flow. The pump
operates at full speed, 100% of the time, at the P2 pressure and design flow,
as shown in Figure 6.17.
In the second situation, the balancing valve is removed and the pump is
operated by a variable-frequency drive at a constant reduced speed, adjusted
manually (or by a flowmeter as in Figure 6.16) to achieve the required system 100% design flow all of the time. This results in pump operation at the
intersection of P3 and the design flow, as shown in Figure 6.17. The result is
an annual energy savings of 33,969 kW-h with the adjustable frequency drive
compared with the balancing valve, as shown in Figure 6.18.
In the last comparison, a variable-frequency drive operates in a closed-loop
control based on the system load, by controlling temperature in the cooling
tower basin. The system load profile is shown in Figure 6.15. A minimum
drive output to maintain at least 60% flow has been applied. The resulting
variable-speed operation saves 171,059 kW-h annually compared with the
Cooling tower
Chiller
Water
inlet
Basin
Water
outlet
Condenser
water
pump
Flow
meter
FIGURE 6.16
Condenser pump system with an adjustable frequency drive. (Courtesy of Danfoss A/S,
Milwaukee, WI.)
189
Energy Conservation and Life-Cycle Costs
Pressure absorbed by
the throttling valve
S1, S2 = System curves
Pressure
S2
P2
S1
Pump curves
P1
P3
Design
flow
Flow
Flow 1
FIGURE 6.17
Throttling valve energy loss compared with variable-speed pumping. (Courtesy of Danfoss
A/S, Milwaukee, WI.)
balancing valve. Energy savings is better than 50%, even maintaining a 60%
minimum speed.
In addition to the energy savings associated with using variable-speed
pumping, there are other important benefits related to the health of the pump.
A pump operating at reduced speed to achieve a lower flow rate produces
less head and operates closer to its BEP than the alternative of throttling a
constant-speed pump. The slower-speed pump produces lower radial and
axial bearing loads, which should give longer bearing life and less chance
Configuration
Discharge balancing
valve, full speed pump
Adjustable frequency
drive, reduced
constant speed pump
Adjustable frequency
drive at variable
speed in closed loop
control
Totals
% Flow
% Hours
Run hours
Power kW
Energy kWh
100
100
8760
30
298,483
100
60
70
80
90
100
100
20
25
35
15
5
100%
8760
1752
2190
3066
1314
438
8760
25.22
5.45
8.65
12.91
18.38
25.22
264,514
12,613
23,948
48,542
29,096
1326
127,424
FIGURE 6.18
Energy savings example. (Courtesy of Danfoss A/S, Milwaukee, WI.)
190
Pump Characteristics and Applications
of bearing or seal failure. Seal face wear should be less on the slower-speed
pump. The pump operating closer to its BEP should produce less internal
recirculation, which can cause erosive damage to the pump. Further, the
pump is less likely to experience that instability sometimes associated with
operating too far to the left or right of BEP.
Variable-speed pumping permits the continued ability to “tweak” a system so that the pump is performing optimally, despite changes in either the
system or the pump. If head requirements change over time, for example,
due to build-up of corrosion products in a pipe line, the pump can simply be
speeded up to account for the higher head required. Or, as the pump impeller wears and clearances open up, the speed can be adjusted upward to keep
the pump from delivering less flow. The fudge factors often used in pump
head calculations (and which may result in the oversizing of the pump as
described in Chapter 2, Section XII) are essentially eliminated if variablespeed pumping is used.
The final advantage of variable speed pumping has to do with the possible reduction of electric power costs due to reduced-speed starting of the
pump. For some high-energy pumps, the high amp draw when the motor
is started at full speed can actually affect the overall rate that the electric
utility charges the user. By starting the pump at a slower speed, the cost per
unit of electric power may be reduced. This concept of using a VFD or other
variable-speed device to soft start the pump is sometimes justification alone
for a variable-speed system.
In summary, variable-speed pumping has been shown to be a highly effective way to reduce total pumping costs for systems that require a wide range
of pump flow. It is used as an alternative to throttling of a single pump, or
the use of multiple pumps in the system. Use of variable-frequency drives
(Figure 6.19) is considered the most likely alternative to achieve meaningful
savings. The advantages of using VFDs to achieve variable-speed pumping
include
• Energy savings
• Ability to sometimes retrofit existing equipment without buying
new motors
• Wide speed change achievable without serious energy loss conse­
quences
• Ability to fine-tune speed in response to changes in the system or
pump
• Lower bearing loads and overall better health of the pump
• Ability to start the pump at reduced speed
• Ability to program the VFD to not dwell in a speed range that is too
close to the natural frequency of the pump and motor, which could
result in resonance and high vibration levels
Energy Conservation and Life-Cycle Costs
191
FIGURE 6.19
Variable-frequency drive. (Courtesy of Danfoss Graham, Milwaukee, WI.)
Surveys of pump users and engineers who design pumping systems using
variable speed indicate that there are circumstances when applications for
pumps as small as 5 to 10 HP can justify the addition of a variable-speed device.
The software introduced in Chapter 3, Section III is an excellent tool for
evaluating annual power savings achievable by using variable-speed pumping and for comparing this option to alternatives of fixed speed with throttled flow or the use of multiple pumps in the system.
V. Pump Life-Cycle Costs
This section is printed courtesy of and with joint permission of the
Hydraulic Institute, Parsippany, NJ; Europump, Brussels, Belgium; and the
U.S. Department of Energy’s Office of Industrial Technologies, Washington,
DC. The material is an executive summary of the 194-page book, Pump Life
Cycle Costs: A Guide to LCC Analysis for Pumping Systems, published by the
Hydraulic Institute and Europump. For more information on this publication, contact the Hydraulic Institute at 973-267-9700 or visit www.pumps.org
or Europump at 32-2-706-8230 or visit www.europump.org.
192
Pump Characteristics and Applications
Note that the LCC example given at the end of this section includes cost
data given in both euros and USD, without adjusting for the relative values
of these two currencies.
A. Improving Pump System Performance: An Overlooked Opportunity?
Pumping systems account for nearly 20% of the world’s electrical energy
demand and range from 25% to 50% of the energy usage in certain industrial
plant operations (Figure 6.20). Pumping systems are widespread; they provide
domestic services, commercial, and agricultural services, municipal water/
wastewater services, and industrial services for food processing, chemical,
petrochemical, pharmaceutical, and mechanical industries. Although pumps
are typically purchased as individual components, they provide a service only
when operating as part of a system. The energy and materials used by a system depend on the design of the pump, the design of the installation, and the
way the system is operated. These factors are interdependent. What is more,
they must be carefully matched to each other, and remain so throughout their
working lives to ensure the lowest energy, and maintenance costs, equipment
life, and other benefits. The initial purchase price is a small part of the LCC for
high usage pumps. Although operating requirements may sometimes override energy cost considerations, an optimum solution is still possible.
FIGURE 6.20
In some industrial plant operations, pumping systems account for 25 to 50% of energy use.
(Courtesy of and with joint permission of the Hydraulic Institute, Parsippany, NJ; Europump,
Brussels, Belgium; the U.S. Department of Energy Office of Industrial Technologies,
Washington, DC.)
193
Energy Conservation and Life-Cycle Costs
A greater understanding of all the components that make up the total cost
of ownership will provide an opportunity to dramatically reduce energy,
operational, and maintenance costs. Reducing energy consumption and
waste also has important environmental benefits.
LCC analysis is a management tool that can help companies minimize
waste and maximize energy efficiency for many types of systems, including pumping systems. This overview provides highlights from Pump Life
Cycle Costs: A Guide to LCC Analysis for Pumping Systems, developed by the
Hydraulic Institute and Europump to assist plant owners/operators in applying the LCC methodology to pumping systems. For information on obtaining a copy of the Guide, see Section V.J.
B. What Is Life-Cycle Cost?
The LCC of any piece of equipment is the total “lifetime” cost to purchase,
install, operate, maintain, and dispose of that equipment. Determining LCC
involves following a methodology to identify and quantify all of the components of the LCC equation.
When used as a comparison tool between possible design or overhaul alternatives, the LCC process will show the most cost-effective solution within
the limits of the available data.
The components of an LCC analysis typically include initial costs, installation and commissioning costs, energy costs, operation costs, maintenance
and repair costs, downtime costs, environmental costs, and decommissioning and disposal costs (see Figure 6.21).
C. Why Should Organizations Care about Life-Cycle Cost?
Many organizations only consider the initial purchase and installation
cost of a system. It is in the fundamental interest of the plant designer
or manager to evaluate the LCC of different solutions before installing
Initial costs
Maintenance
costs
Energy costs
Other costs
FIGURE 6.21
Typical LCCs for a medium-sized industrial plant. (Courtesy of and with joint permission of
the Hydraulic Institute, Parsippany, NJ; Europump, Brussels, Belgium; the U.S. Department of
Energy Office of Industrial Technologies, Washington, DC.)
194
Pump Characteristics and Applications
major new equipment or carrying out a major overhaul. This evaluation
will identify the most financially attractive alternative. As national and
global markets continue to become more competitive, organizations must
continually seek cost savings that will improve the profitability of their
operations. Plant equipment operations are receiving particular attention
as a source of cost savings, especially minimizing energy consumption
and plant downtime.
Existing systems provide a greater opportunity for savings through the
use of LCC methods than do new systems for two reasons. First, there are at
least 20 times as many pump systems in the installed base as are built each
year; second, many of the existing systems have pumps or controls that are
not optimized since the pumping tasks change over time.
Some studies have shown that 30% to 50% of the energy consumed
by pump systems could be saved through equipment or control system
changes.
In addition to the economic reasons for using LCC, many organizations
are becoming increasingly aware of the environmental impact of their businesses, and are considering energy efficiency as one way to reduce emissions
and preserve natural resources.
D. Getting Started
LCC analysis, either for new facilities or renovations, requires the evaluation
of alternative systems. For a majority of facilities, the lifetime energy and/or
maintenance costs will dominate the LCCs. It is therefore important to accurately determine the current cost of energy, the expected annual energy price
escalation for the estimated life, along with the expected maintenance labor
and material costs. Other elements, such as the lifetime costs of downtime,
decommissioning, and environmental protection, can often be estimated
based on historical data for the facility. Depending upon the process, downtime costs can be more significant than the energy or maintenance elements
of the equation. Careful consideration should therefore be given to productivity losses due to downtime.
This overview provides an introduction to the life-cycle costing process.
The complete Guide expands upon life-cycle costing and provides substantial
technical guidance on designing new pumping systems as well as assessing
improvements to existing systems. The Guide also includes a sample chart,
examples of manual calculation of LCC, and a software tool to assist in LCC
calculation.
E. Life-Cycle Cost Analysis
In applying the evaluation process or in selecting pumps and other equipment, the best information concerning the output and operation of the
plant must be established. The process itself is mathematically sound, but
Energy Conservation and Life-Cycle Costs
195
if incorrect or imprecise information is used, then an incorrect or imprecise
assessment will result. The LCC process is a way to predict the most costeffective solution; it does not guarantee a particular result, but allows the
plant designer or manager to make a reasonable comparison between alternate solutions within the limits of the available data.
Pumping systems often have a lifespan of 15 to 20 years. Some cost elements will be incurred at the outset, and others may be incurred at different
times throughout the lives of the different solutions being evaluated. It is
therefore practicable, and possibly essential, to calculate a present or discounted value of the LCC to accurately assess the different solutions.
This analysis is concerned with assessments where details of the system
design are being reviewed. Here the comparison is between one pump type
and another or one control means and another. The exercise may be aimed
at determining what scope could be justified for a monitoring or control
scheme, or for different process control means to be provided. Whatever the
specifics, the designs should be compared on a like-for-like basis. To make
a fair comparison, the plant designer/manager might need to consider the
measure used. For example, the same process output volume should be considered and, if the two items being examined cannot give the same output
volume, it may be appropriate to express the figures in cost per unit of output
(e.g., $/ton or euros/kg). The analysis should consider all significant differences between the solutions being evaluated.
Finally, the plant designer or manager might need to consider maintenance or servicing costs, particularly where these are to be subcontracted,
or spare parts are to be provided with the initial supply of the equipment for
emergency standby provision. Whatever is considered must be on a strictly
comparable basis. If the plant designer or manager decides to subcontract
or carry strategic spares based entirely on the grounds of convenience, this
criterion must be used for all systems being assessed. But, if it is the result of
maintenance that can be carried out only by a specialist subcontractor, then
its cost will correctly appear against the evaluation of that system.
Elements of the LCC equation are as follows:
LCC = Cic + Cin + Ce + Co + Cm + Cs + Cenv + Cd
(6.5)
where LCC is the life-cycle cost, Cic is the initial costs, purchase price (pump,
system, pipe, auxiliary services), Cin is the installation and commissioning
costs (including training), Ce is the energy costs (predicted cost for system
operation, including pump driver, controls, and any auxiliary services), Co is
the operation costs (labor cost of normal system supervision), Cm is the maintenance and repair costs (routine and predicted repairs), Cs is the downtime
costs (loss of production), Cenv is the environmental costs (contamination
from pumped liquid and auxiliary equipment), Cd is the decommissioning/
disposal costs (including restoration of the local environment and disposal
of auxiliary services).
196
Pump Characteristics and Applications
The following sections examine each element and offer suggestions on
how a realistic value can be determined for use in computing the LCC. It
should be noted that this calculation does not include the raw materials consumed by the plant in making a product.
1. Cic—Initial Investment Costs
The pump plant designer or manager must decide the outline design of the
pumping system. The smaller the pipe and fitting diameters, the lower the cost
of acquiring and installing them. However, the smaller diameter installation
requires a more powerful pump resulting in higher initial and operating
costs. In addition, smaller pipe sizes on the inlet side of a pump will reduce
the net positive suction head available (NPSHa), thus requiring a larger and
slower-speed pump, which will typically be more expensive. Provisions
must be made for the acceleration head needed for positive displacement
pumps or the depth of submergence needed for a wet pit pump.
There will be other choices, which may be made during the design stage,
that can affect initial investment costs. One important choice is the quality
of the equipment being selected. There may be an option regarding materials having differing wear rates, heavier duty bearings, or seals, or more
extensive control packages, all increasing the working life of the pump.
These and other choices may incur higher initial costs but reduce LCC
costs.
The initial costs will also usually include the following items:
• Engineering (e.g., design and drawings, regulatory issues)
• Bid process
• Purchase order administration
• Testing and inspection
• Inventory of spare parts
• Training
• Auxiliary equipment for cooling and sealing water
2. Cin—Installation and Commissioning (Start-Up) Costs
Installation and commissioning costs include the following:
• Foundations—design, preparation, concrete and reinforcing, etc.
• Setting and grouting of equipment on foundation
• Connection of process piping
• Connection of electrical wiring and instrumentation
Energy Conservation and Life-Cycle Costs
197
• Connection of auxiliary systems and other utilities
• Provisions for flushing or “water runs”
• Performance evaluation at startup
Installation can be accomplished by an equipment supplier, contractor,
or by user personnel. This decision depends on several factors, including the skills, tools, and equipment required to complete the installation;
contractual procurement requirements; work rules governing the installation site; and the availability of competent installation personnel. Plant
or contractor personnel should coordinate site supervision with the supplier. Care should be taken to follow installation instructions carefully. A
complete installation includes transfer of equipment operation and maintenance requirements via training of personnel responsible for system
operation.
Commissioning requires close attention to the equipment manufacturer’s
instruction for initial startup and operation. A checklist should be used
to ensure that equipment and the system are operating within specified
parameters. A final sign-off typically occurs after successful operation is
demonstrated.
3. Ce—Energy Costs
Energy consumption is often one of the larger cost elements and may dominate the LCC, especially if pumps run more than 2000 h per year. Energy
consumption is calculated by gathering data on the pattern of the system
output. If output is steady, or essentially so, the calculation is simple. If output varies over time, then a time-based usage pattern needs to be established.
The input power calculation formulae are
P = (Q × H × SG)/(366 × ηp × ηm) [kW] (metric)
(6.6)
P = (Q × H × SG)/(3960 × ηp × ηm) [hp] (U.S. units)
(6.7)
where P is the power, Q is the rate of flow, m3/h (U.S. gpm), H is the head, m
(ft), ηp is the pump efficiency, ηm is the motor efficiency, and SG is the specific
gravity.
The plant designer or manager needs to obtain separate data showing
the performance of each pump/system being considered over the output
range. Performance can be measured in terms of the overall efficiencies of
the pump unit or of the energies used by the system at the different output
levels. Driver selection and application will affect energy consumption. For
example, much more electricity is required to drive a pump with an air
motor than with an electric motor. In addition, some energy use may not be
198
Pump Characteristics and Applications
output dependent. For example, a control system sensing output changes
may itself generate a constant energy load, whereas a variable-speed electric motor drive may consume different levels of energy at different operating settings. The use of a throttling valve, pressure relief, or flow by-pass
for control will reduce the operating efficiency and increase the energy
consumed.
The efficiency or levels of energy used should be plotted on the same time
base as the usage values to show their relationship to the usage pattern. The
area under the curve then represents the total energy absorbed by the system being reviewed over the selected operating cycle. The result will be in
kilowatt-hours (kW-h). If there are differential power costs at different levels
of load, then the areas must be totaled within these levels.
Once the charge rates are determined for the energy supplied, they can be
applied to the total kW-h for each charge band (rate period). The total cost
of the energy absorbed can then be found for each system under review and
brought to a common time period.
Finally, the energy and material consumption costs of auxiliary services
need to be included. These costs may come from cooling or heating circuits,
from liquid flush lines, or liquid/gas barrier arrangements. For example, the
cost of running a cooling circuit using water will need to include the following items: cost of the water, booster pump service, filtration, circulation, and
heat extraction/dissipation.
4. Co—Operation Costs
Operation costs are labor costs related to the operation of a pumping system. These vary widely, depending on the complexity and duty of the
system. For example, a hazardous duty pump may require daily checks
for hazardous emissions, operational reliability, and performance within
accepted parameters. On the other hand, a fully automated nonhazardous
system may require very limited supervision. Regular observation of how
a pumping system is functioning can alert operators to potential losses in
system performance. Performance indicators include changes in vibration,
shock pulse signature, temperature, noise, power consumption, flow rates,
and pressure.
5. Cm—Maintenance and Repair Costs
Obtaining optimum working life from a pump (Figure 6.22) requires regular
and efficient servicing. The manufacturer will advise the user about the frequency and the extent of this routine maintenance. Its cost depends on the
time and frequency of service and the cost of materials. The design can influence these costs through the materials of construction, components chosen,
and the ease of access to the parts to be serviced.
Energy Conservation and Life-Cycle Costs
199
FIGURE 6.22
Maintenance and repair are significant components of pumping system LCCs, and an effective
maintenance program can minimize these costs. (Courtesy of and with joint permission of
the Hydraulic Institute, Parsippany, NJ; Europump, Brussels, Belgium; the U.S. Department of
Energy Office of Industrial Technologies, Washington, DC.)
The maintenance program can comprise less frequent but more major
attention as well as the more frequent but simpler servicing. The major
activities often require removing the pump to a workshop. During the
time the unit is unavailable to the process plant, there can be loss of product or a cost from a temporary replacement. These costs can be minimized
by programming major maintenance during annual shutdown or process
changeover. Major service can be described as “pump unit not repairable
in place,” whereas the routine work is described as “pump unit repairable
in place.”
The total cost of routine maintenance is found by multiplying the costs per
event by the number of events expected during the life cycle of the pump.
Although unexpected failures cannot be predicted precisely, they can be
estimated statistically by calculating mean time between failures (MTBF).
MTBF can be estimated for components and then combined to give a value
for the complete machine.
It might be sufficient to simply consider best- and worst-case scenarios
where the shortest likely life and the longest likely lifetimes are considered.
In many cases, plant historical data are available.
The manufacturer can define and provide MTBF of the items whose failure
will prevent the pump unit from operating or will reduce its life expectancy
below the design target. These values can be derived from past experience
or from theoretical analyses. The items can be expected to include seals,
200
Pump Characteristics and Applications
bearings, impeller/valve/port wear, coupling wear, motor features, and
other special items that make up the complete system. The MTBF values can
be compared with the design working life of the unit and the number of
failure events calculated.
It must be recognized that process variations and user practices will almost
certainly have a major impact upon the MTBF of a plant and the pumps
incorporated in it. Whenever available, historical data are preferable to theoretical data from the equipment supplier. The cost of each event and the total
costs of these unexpected failures can be estimated in the same way that
routine maintenance costs are calculated.
6. Cs—Downtime and Loss of Production Costs
The cost of unexpected downtime and lost production is a very significant
item in the total LCC and can rival the energy costs and replacement parts
costs in its impact. Despite the design or target life of a pump and its components, there will be occasions when an unexpected failure occurs. In those
cases where the cost of lost production is unacceptably high, a spare pump
may be installed in parallel to reduce the risk. If a spare pump is used, the
initial cost will be greater but the cost of unscheduled maintenance will
include only the cost of the repair.
The cost of lost production is dependent on downtime and differs from
case to case.
7. Cenv—Environmental Costs, Including Disposal of Parts
and Contamination from Pumped Liquid
The cost of contaminant disposal during the lifetime of the pumping system varies significantly, depending on the nature of the pumped product.
Certain choices can significantly reduce the amount of contamination, but
usually at an increased investment cost. Examples of environmental contamination can include cooling water and packing box leakage disposal,
hazardous pumped product flare-off, used lubricant disposal, and contaminated used parts such as seals. Costs for environmental inspection
should also be included.
8. Cd—Decommissioning/Disposal Costs, Including
Restoration of the Local Environment
In the vast majority of cases, the cost of disposing of a pumping system
will vary little with different designs. This is certainly true for nonhazardous liquids and, in most cases, for hazardous liquids also. Toxic, radioactive, or other hazardous liquids will have legally imposed protection
requirements, which will be largely the same for all system designs. A
Energy Conservation and Life-Cycle Costs
201
difference may occur when one system has the disposal arrangements as
part of its operating arrangements (for example, a hygienic pump designed
for cleaning in place), whereas another does not (for example, a hygienic
pump designed for removal before cleaning). Similar arguments can be
applied to the costs of restoring the local environment. When disposal is
very expensive, the LCC becomes much more sensitive to the useful life of
the equipment.
F. Total Life-Cycle Costs
The costs estimated for the various elements making up the total LCCs need
to be aggregated to allow a comparison of the designs being considered. This
is best done by means of a tabulation that identifies each item and asks for
a value to be inserted. Where no value is entered, an explanatory comment
should be added. The estimated costs can then be totaled to give the LCC
values for comparison, and attention will also be drawn to nonqualitative
evaluation factors.
There are also financial factors to take into consideration in developing the
LCC. These include:
• Present energy prices
• Expected annual energy price increase (inflation) during the pumping system life time
• Discount rate
• Interest rate
• Expected equipment life (calculation period)
In addition, the user must decide which costs to include, such as maintenance, downtime, environmental, disposal, and other important costs.
For the calculation of the present worth of a single cost element, refer to
Table 6.3 for present worth factor Cp/Cn, where Cp is the present cost of a
single cost element and Cn is the cost paid after “n” years. For the calculation
of the present worth of constant yearly expenditures, refer to Table 6.4 for the
discount factor that applies for a particular real interest rate and number of
years.
G. Pumping System Design
Proper pumping system design is the most important single element in minimizing the LCC. All pumping systems are comprised of a pump, a driver,
pipe installation, and operating controls, and each of these elements is considered individually. Proper design considers the interaction between the
pump and the rest of the system and the calculation of the operating duty
1
2
3
4
5
6
7
8
9
10
15
20
25
30
–1
1.01
1.02
1.03
1.04
1.05
1.06
1.07
1.08
1.09
1.10
1.15
1.20
1.25
1.30
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
1
0.99
0.98
0.97
0.96
0.95
0.94
0.94
0.93
0.92
0.91
0.87
0.83
0.80
0.77
0.98
0.96
0.94
0.93
0.91
0.89
0.87
0.86
0.84
0.83
0.76
0.69
0.64
0.59
2
0.97
0.94
0.92
0.89
0.86
0.84
0.82
0.79
0.77
0.75
0.66
0.58
0.51
0.46
3
0.96
0.92
0.89
0.86
0.82
0.79
0.76
0.74
0.71
0.68
0.57
0.48
0.41
0.35
4
0.95
0.91
0.86
0.82
0.78
0.75
0.71
0.68
0.65
0.62
0.50
0.40
0.33
0.27
5
0.94
0.89
0.84
0.79
0.75
0.71
0.67
0.63
0.60
0.56
0.43
0.34
0.26
0.21
6
7
0.93
0.87
0.81
0.76
0.71
0.67
0.62
0.58
0.55
0.51
0.38
0.28
0.21
0.16
Real Discount Rate (interest rate minus inflation rate, in %)
8
0.93
0.85
0.79
0.73
0.68
0.63
0.58
0.54
0.50
0.47
0.33
0.23
0.17
0.12
Source: Courtesy of Hydraulic Institute, Parsippany, NJ, www.pumps.org; Europump, Brussels, Belgium, www.europump.org.
–2
1.02
1.04
1.06
1.08
1.10
1.12
1.15
1.17
1.19
1.21
1.32
1.44
1.56
1.69
No. of
Years (n)
Factor Cp/Cn for a Single-Cost Element after n Years
TABLE 6.3
9
0.92
0.84
0.77
0.70
0.65
0.59
0.54
0.50
0.46
0.42
0.28
0.19
0.13
0.09
10
0.91
0.82
0.74
0.68
0.61
0.56
0.51
0.46
0.42
0.39
0.25
0.16
0.11
0.07
202
Pump Characteristics and Applications
1.02
2.06
3.12
4.21
5.31
6.44
7.60
8.77
9.97
11.19
17.20
24.89
32.85
41.66
–2
1.01
2.03
3.06
4.10
5.15
6.22
7.29
8.37
9.47
10.57
16.27
22.26
28.56
35.19
–1
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
15.00
20.00
25.00
30.00
0
0.99
1.97
2.94
3.90
4.85
5.80
6.73
7.65
8.57
9.47
13.87
18.05
22.02
25.81
1
0.98
1.94
2.88
3.81
4.71
5.60
6.47
7.33
8.16
8.98
12.85
16.35
19.52
22.40
2
0.97
1.91
2.83
3.72
4.58
5.42
6.23
7.02
7.79
8.53
11.94
14.88
17.41
19.60
3
0.96
1.89
2.78
3.63
4.45
5.24
6.00
6.73
7.44
8.11
11.12
13.59
15.62
17.29
4
0.95
1.86
2.72
3.55
4.33
5.08
5.79
6.46
7.11
7.72
10.38
12.46
14.09
15.37
5
0.94
1.83
2.67
3.47
4.21
4.92
5.58
6.21
6.80
7.36
9.71
11.47
12.78
13.76
6
0.93
1.81
2.62
3.39
4.10
4.77
5.39
5.97
6.52
7.02
9.11
10.59
11.65
12.41
7
Real Discount Rate (interest rate minus inflation rate, in percent)
0.93
1.78
2.58
3.31
3.99
4.62
5.21
5.75
6.25
6.71
8.56
9.82
10.67
11.26
8
Source: Courtesy of Hydraulic Institute, Parsippany, NJ, www.pumps.org; and Europump, Brussels, Belgium, www.europump.org.
1
2
3
4
5
6
7
8
9
10
15
20
25
30
No. of
Years (n)
Discount Factor (df) for Constant Yearly Expenditures
TABLE 6.4
0.92
1.76
2.53
3.24
3.89
4.49
5.03
5.53
6.00
6.42
8.06
9.13
9.82
10.27
9
0.91
1.74
2.49
3.17
3.79
4.36
4.87
5.33
5.76
6.14
7.61
8.51
9.08
9.43
10
Energy Conservation and Life-Cycle Costs
203
204
Pump Characteristics and Applications
point(s). The characteristics of the piping system must be calculated to determine required pump performance. This applies to both simple systems as
well as to more complex (branched) systems.
Both procurement costs and operational costs make up the total cost of
an installation during its lifetime. A number of installation and operational
costs are directly dependent on the piping diameter and the components in
the piping system.
A considerable amount of the pressure losses in the system is caused by
valves, in particular control valves in throttle-regulated installations. In
systems with several pumps, the pump workload is divided between the
pumps, which together, and in conjunction with the piping system, deliver
the required flow.
The piping diameter is selected based on the following factors:
• Economy of the whole installation (pumps and system)
• Required lowest flow velocity for the application (e.g., avoid
sedimentation)
• Required minimum internal diameter for the application (e.g., solids
handling)
• Maximum flow velocity to minimize erosion in piping and fittings
• Plant standard pipe diameters
Decreasing the pipeline diameter has the following effects:
• Piping and component procurement and installation costs will decrease.
• Pump installation procurement costs will increase as a result of
increased flow losses with consequent requirements for higher head
pumps and larger motors. Costs for electrical supply systems will
therefore increase.
• Operating costs will increase as a result of higher energy usage due
to increased friction losses.
Some costs increase with increasing pipeline size and some decrease.
Because of this, an optimum pipeline size may be found, based on minimizing costs over the life of the system.
The duty point of the pump is determined by the intersection of the system
curve and the pump curve as shown in Figure 6.23.
A pump application might need to cover several duty points, of which the
largest flow and/or head will determine the rated duty for the pump. The
pump user must carefully consider the duration of operation at the individual duty points to properly select the number of pumps in the installation
and to select output control. Many software packages are currently available
that make it easier to determine friction losses and generate system curves
205
Energy Conservation and Life-Cycle Costs
H
(head)
Pump curve
System curve
Duty point
Static
head
Q (rate of flow)
FIGURE 6.23
The duty point is the intersection between the pump and system curves. (Courtesy of and with
joint permission of the Hydraulic Institute, Parsippany, NJ; Europump, Brussels, Belgium; the
U.S. Department of Energy Office of Industrial Technologies, Washington, DC.)
(see Chapter 3, Section III). Most pump manufacturers can recommend software suitable for the intended duty. Different programs may use different
methods of predicting friction losses and may give slightly different results.
Very often, such software is also linked to pump-selection software from
that particular manufacturer.
H. Methods for Analyzing Existing Pumping Systems
The following steps provide an overall guideline to improve an existing
pumping system.
• Assemble a complete document inventory of the items in the pumping system.
• Determine the flow rates required for each load in the system.
• Balance the system to meet the required flow rates of each load.
• Minimize system losses needed to balance the flow rates.
• Effect changes to the pump to minimize excessive pump head in the
balanced system.
• Identify pumps with high maintenance cost.
One of two methods can be used to analyze existing pumping systems.
One consists of observing the operation of the actual piping system, and
206
Pump Characteristics and Applications
the second consists of performing detailed calculations using fluid analysis techniques. The first method relies on observations of the operating piping system (pressures, differential pressures, and flow rates); the
second deals with creating an accurate mathematical model of the piping system and then calculating the pressures and flow rates within the
model.
Observing the operating system allows one to view how the actual system is working, but system operational requirements limit the amount of
experimentation that plant management will allow. By developing a model
of the piping system, one can easily consider system alternatives, but the
model must first be validated to ensure that it accurately represents the operating piping system it is trying to emulate. Regardless of the method used,
the objective is to gain a clear picture of how the various parts of the system operate and to see where improvements can be made and the system
optimized.
The following is a checklist of some useful means to reduce the LCC of a
pumping system.
• Consider all relevant costs to determine the LCC.
• Procure pumps and systems using LCC considerations.
• Optimize total cost by considering operational costs and procurement costs.
• Consider the duration of the individual pump duty points.
• Match the equipment to the system needs for maximum benefit.
• Match the pump type to the intended duty.
• Do not oversize the pump.
• Match the driver type to the intended duty.
• Specify motors to be high efficiency.
• Match the power transmission equipment to the intended duty.
• Evaluate system effectiveness.
• Monitor and sustain the pump and system to maximize benefit.
• Consider the energy wasted using control valves.
• Utilize auxiliary services wisely.
• Optimize preventative maintenance.
• Maintain the internal pump clearances.
• Follow available guidelines regarding the rewinding of motors.
• Analyze existing pump systems for improvement opportunities.
• Use the showcases in the Guide as a source for ideas.
207
Energy Conservation and Life-Cycle Costs
I. Example: Pumping System with a Problem Control Valve
In this example, the LCC analysis for the piping system is directed at a
control valve. The system is a single pump circuit that transports a process
fluid containing some solids from a storage tank to a pressurized tank
(Figure 6.24). A heat exchanger heats the fluid, and a control valve regulates the rate of flow into the pressurized tank to 80 m3/h (350 gal/min
[gpm]).
The plant engineer is experiencing problems with an FCV that fails due
to erosion caused by cavitation. The valve fails every 10 to 12 months at
a cost of 4000 euros or USD per repair. A change in the control valve is
being considered to replace the existing valve with one that can resist cavitation. Before changing out the control valve again, the project engineer
wanted to look at other options and perform a LCC analysis on alternative
solutions.
The first step is to determine how the system is currently operating and
determine why the control valve fails, then to see what can be done to correct
the problem.
The control valve currently operates between 15% and 20% open and with
considerable cavitation noise from the valve. It appears the valve was not
sized properly for the application. After reviewing the original design calculations, it was discovered that the pump was oversized; 110 m3/h (485 gpm)
Pressure tank
2.0 bar
Storage tank
Pump
FCV at 15%
Heat exchanger
FIGURE 6.24
Sketch of pumping system in which the control valve fails. (Courtesy of and with joint permission of the Hydraulic Institute, Parsippany, NJ; Europump, Brussels, Belgium; the U.S.
Department of Energy Office of Industrial Technologies, Washington, DC.)
208
Pump Characteristics and Applications
instead of 80 m3/h (350 gpm). This resulted in a larger pressure drop across
the control valve than originally intended.
As a result of the large differential pressure at the operating rate of
flow, and the fact that the valve is showing cavitation damage at regular
intervals, it is determined that the control valve is not suitable for this
process.
The following four options are suggested:
1. A new control valve can be installed to accommodate the high pressure differential.
2. The pump impeller can be trimmed so that the pump does not
develop as much head, resulting in a lower pressure drop across the
current valve.
3. A variable frequency drive (VFD) can be installed, and the FCV
removed. The VFD can vary the pump speed and thus achieve the
desired process flow.
4. The system can be left as it is, with a yearly repair of the FCV to be
expected.
The cost of a new control valve that is properly sized is 5000 euros or USD.
The cost of modifying the pump performance by reduction of the impeller
diameter is 2250 euros or USD. The process operates at 80 m3/h for 6000 h/
year. The energy cost is 0.08 euro or USD per kW-h and the motor efficiency
is 90%.
The cost comparison of the pump system modification options is contained in Table 6.5. Figure 6.25 shows the pump and system curves showing
the operation of the original system and the modified impeller.
By trimming the impeller to 375 mm (Option 2), the pump’s total head
is reduced to 42.0 m (138 ft) at 80 m3/h. This drop in pressure reduces the
differential pressure across the control valve to less than 10 m (33 ft), which
better matches the valve’s original design point. The resulting annual energy
cost with the smaller impeller is 6720 euros or USD per year. It costs 2250
euros or USD to trim the impeller. This includes the machining cost as well
as the cost to disassemble and reassemble the pump.
A 30-kW VFD (Option 3) costs 20,000 euro or USD and an additional 1500
euros or USD to install. The VFD will cost 500 euros or USD to maintain each
year. It is assumed that it will not need any repairs over the project’s 8-year
life.
The option to leave the system unchanged (Option 4) will result in a yearly
cost of 4000 euros or USD for repairs to the cavitating FCV.
LCC costs and assumptions are as follows:
• The current energy price is 0.08 euro or USD per kW-h.
• The process is operated for 6000 h/year.
430 mm
71.7 m (235 ft)
75.1%
80 m3/h (350 US gpm)
23.1 kW
11,088 euros or USD
5000 euros or USD
0
0
0
0
Cost
Pump Cost Data
Impeller diameter
Pump head
Pump efficiency
Rate of flow
Power consumed
Energy cost/year
New valve
Modify impeller
VFD
Installation of VFD
Valve repair/year
375 mm
42.0 m (138 ft)
72.1%
80 m3/h (350 US gpm)
14.0 kW
6720 euros or USD
0
2250 euros or USD
0
0
0
Option 2
Trim Impeller
430 mm
34.5 m (113 ft)
77%
80 m3/h (350 US gpm)
11.6 kW
5568 euros or USD
0
0
20,000 euros or USD
1500 euros or USD
0
Option 3
VFD and Remove
Control Valve
4000 euros or USD
430 mm
71.7 m (235 ft)
75.1%
80 m3/h (350 US gpm)
23.1 kW
11,088 euros or USD
0
0
0
Option 4
Repair Control Valve
Source: Courtesy of and with joint permission of the Hydraulic Institute, Parsippany, NJ; Europump, Brussels, Belgium; the U.S. Department of Energy
Office of Industrial Technologies, Washington, D.C.
Option 1
Change Control Valve
Cost Comparison for Options 1 through 4 in the System with a Failing Control Valve System
TABLE 6.5
Energy Conservation and Life-Cycle Costs
209
210
Pump Characteristics and Applications
100
“System curve” option 1 and 4
“System curve” option 2
Head (m)
80
430 mm impeller
100
“System curve” option 3
80
60
40
20
0
% Efficiency
375 mm impeller
60
40
Reduced speed pump curve
0
20
40
60
80
100
120
140
160
20
Rate of flow (m3/h)
FIGURE 6.25
Pump and system curves showing the operation of the original system and the modified pump
impeller. (Courtesy of and with joint permission of the Hydraulic Institute, Parsippany, NJ;
Europump, Brussels, Belgium; the U.S. Department of Energy Office of Industrial Technologies,
Washington, DC.)
• The company has an annual cost for routine maintenance for pumps
of this size at 500 euros or USD per year, with a repair cost of 2500
euros or USD every second year.
• There is no decommissioning cost or environmental disposal cost
associated with this project.
• This project has an 8-year life.
• The interest rate for new capital projects is 8%, and an inflation rate
of 4% is expected.
The LCC calculations for each of the four options are summarized in
Table 6.6. Option 2, trimming the impeller, has the lowest LCC, and is the
preferred option for this example.
J. For More Information
To order Pump Life Cycle Costs: A Guide to LCC Analysis for Pumping Systems,
contact the Hydraulic Institute or Europump.
2250
0.080
14.0
6000
6720
500
2500
0
0
0
0
8
8.0
4.0
59,481
5000
0.080
23.1
6000
11,088
500
2500
0
0
0
0
8
8.0
4.0
91,827
Option 2
Trim
Impeller
74,313
1000
2500
0
0
0
0
8
8.0
4.0
21,500
0.080
11.6
6000
5568
Option 3
VFD and Remove
Control Valve
113,930
500
2500
4000
0
0
0
8
8.0
4.0
0
0.080
23.1
6000
11,088
Option 4
Repair Control
Valve
Source: Courtesy of and with joint permission of the Hydraulic Institute, Parsippany, NJ; Europump, Brussels, Belgium; and the U.S. Department of
Energy Office of Industrial Technologies, Washington, DC.
Input
Initial investment cost
Energy price (present) per kW-h
Weighted average power of equipment (kW)
Average operating hours/year
Energy cost/year (calculated as energy price ×
weighted average power × average operating
hours/year)
Maintenance cost (routine maintenance)/year
Repair every second year
Other yearly costs
Downtime cost/year
Environmental cost
Decommissioning/disposal (salvage) cost
Lifetime in years
Interest rate (%)
Inflation rate (%)
Output
Present LCC value
Option 1
Change
Control Valve
LCC Comparison for the Problem Control Valve System
TABLE 6.6
Energy Conservation and Life-Cycle Costs
211
212
Pump Characteristics and Applications
1. About the Hydraulic Institute
The Hydraulic Institute (HI), established in 1917, is the largest association of
pump producers and leading suppliers in North America. HI serves member companies and pump users by providing product standards and forums
for the exchange of industry information. HI has been developing pump
standards for over 80 years. For information on membership, organization
structure, member and user services, and energy and LCC issues, visit www.
pumps.org.
Hydraulic Institute
6 Campus Drive – First Floor North
Parsippany, NJ 07054
973-267-9700 (phone)
973-267-9055 (fax)
2. About Europump
Europump, established in 1960, acts as spokesman for 15 national pump
manufacturing associations in Europe, and represents more than 400 manufacturers. Europump serves and promotes the European pump industry.
For information regarding Europump work in the field of LCC issues, please
e-mail secretariat@europump.org. For information on Europump, visit
www.europump.org.
Europump
80 Bd Reyers
1030 Brussels, Belgium
+32 2 706 82 30 (phone)
3. About the U.S. Department of Energy’s
Office of Industrial Technologies
DOE’s Office of Industrial Technologies (OIT) is now renamed Energy
Efficiency and Renewable Energy. Through partnerships with industrial
companies and trade groups, this office develops and delivers advanced
energy efficiency, renewable energy, and pollution prevention technologies
for industrial applications. Visit www.eere.energy.gov or call 800-862-2086 to
learn more about these programs and services.
7
Special Pump-Related Topics
I. Overview
The fundamentals of centrifugal pump design have remained largely
unchanged over the past 50 years. Although CFD analysis is usually
employed today by engineers who do hydraulic design, the design basics
of pump impellers, volutes, and diffusers have not fundamentally changed.
The shapes of pump performance curves and the characteristics of pumps
operating in systems are little different from the way they appeared in
pumps and systems half a century ago. However, the consequences of misapplying pumps (i.e., choosing inappropriate configurations, poor material
selections, incorrect sizing of pumps, or operating pumps too far from BEP
on their H–Q curves) are far more significant than they were 50 years ago.
The reasons for this include the fact that energy is much more expensive,
maintenance of pumps is far more costly, and downtime in a production
facility is a far more expensive thing to consider than was the case 50 years
ago.
The above paragraph is not meant to imply that nothing new has happened in pump technology in the past 50 years. A better understanding of
suction specific speed and hydraulic shaft loads has contributed significantly
to pump reliability. There have been a great many achievements in manufacturing techniques resulting in higher quality and more durable pumps.
Improvements in these areas include better metal casting techniques such as
investment casting, producing cast parts of greater integrity and requiring
less machining and repair; and more precise machine tools and techniques,
producing more accurate fits and closer tolerances.
Many improvements in pump reliability and the ability to handle highly
corrosive and abrasive liquids have been achieved through the development
of superior materials, including both metal alloys and nonmetallic materials.
A number of new pump configurations have come to prominence in the
past 30 years, replacing earlier pump configurations because of higher reliability, lower cost, and/or other benefits. Examples include submersible
pumps supplanting vertical lineshaft or column-type pumps, wet rotor
inline technology being used in place of coupled pumps for residential hot
© 2008 Taylor & Francis Group, LLC
213
214
Pump Characteristics and Applications
water circulators, and air-operated diaphragm pumps replacing other positive displacement or centrifugal alternatives.
There have also been numerous advances in specific aspects of pump
mechanical design. Some examples include the development of better packing and mechanical seal types and materials, a wider variety of bearing
materials and lubrication systems, and material options to reduce abrasive
wear in areas subject to such damage.
This chapter describes a few special topics related to pumps that are considered to have contributed significantly to the broadening of the application
range of pumps and to the improvement in pump reliability. The topics discussed here are not meant to be an all-inclusive listing of all of the technological advances. Some of these topics have been covered in other chapters
of this book as well.
II. Variable-Speed Systems
Variable-speed pumping systems are covered in Chapter 6, Section IV. The
technology has evolved through a number of alternative techniques for
achieving variable-speed operation of pumps. The use of variable-frequency
drives (VFDs) to achieve this end is considered by many to be the best alternative for the broadest range of pump applications. The use of variablespeed technology for those pumping applications where it is appropriate has
resulted in significant energy savings, improved pump performance, and
reduced maintenance costs, compared with constant-speed alternatives.
The growth in applications for VFDs has resulted in a dramatic increase
in the number of suppliers of VFD equipment; improvements in the quality,
reliability, and durability of the products; and a reduction in the cost of the
equipment. There are presently a great many suppliers of variable-frequency
drives in the United States, including virtually all of the major makers of
electric motors and electronic control devices.
III. Sealless Pumps
Sealless pumps are covered in Chapter 5, Section V. Several types of positive
displacement pumps (diaphragm and peristaltic pumps) are discussed as
sealless alternatives in that chapter, as is the rather mundane vertical column
centrifugal pump.
The two primary alternatives for sealless centrifugal pumping—magnetic
drive and canned motor pumps—are compared and contrasted in Chapter 5.
© 2008 Taylor & Francis Group, LLC
Special Pump-Related Topics
215
Both technologies have been around for many years, but both are continuing
to expand to higher horsepower pumps, more aggressive liquids, and higher
operating temperatures.
The need for sealless pumping is continuing to grow as more and more liquids are put into the category for which zero leakage is accepted, and as the
search for greater pump reliability attempts to eliminate one of the leading
contributors to pump downtime, the shaft sealing system.
As Chapter 5 points out, neither the magnetic drive nor the canned motor
design type is a panacea, with the major weakness for both design types
being the fact that radial and thrust loads generated by the pump must be
accommodated by sleeve bearings and thrust plates that are often exposed
to the liquid being pumped. Other limitations include upper limits of viscosity, concerns for inadvertent dry running (especially for mag drive pumps),
special motors required for canned motor pumps, complication of design
compared with sealed pumps, and the resulting high level of maintenance
and reduced reliability that these factors suggest.
Despite these shortcomings, manufacturers are working hard to improve
sealless pump designs, and there is no doubt that sealless technology will
play an important role in the pump industry for the foreseeable future.
IV. Corrosion
Corrosion attack occurs on many components in a plant in addition to pumps,
so many of the comments in this section are relevant to other equipment as
well as pumps. The discussion in this section is restricted to corrosive attack
on metals. The chemistry and nature of corrosion-like attack on nonmetallic
materials is quite different from that for metals. Section V to follow discusses
properties of nonmetallic components used in pumps, whereas Section VI
covers the properties of elastomers used for O-rings in pumps.
When discussing the corrosion of metals, the concept of a local cathode
and anode on a metallic surface provides a general description of how most
corrosive attacks occur. No matter what the type of corrosion in metals, the
basic nature is always the same: a flow of electricity between two areas of
a different electrical potential, through a solution capable of conducting an
electric current. For corrosion to occur in metals, three separate conditions
must occur:
• There must be a difference in electrical potential (i.e., there must be
an anode and a cathode).
• The areas that are different in electrical potential must be in an
electrolyte.
© 2008 Taylor & Francis Group, LLC
216
Pump Characteristics and Applications
• There must be a metallic connection between the areas of potential
difference.
The area of the lower electrical potential is called the anode, or the negative pole, and is the area where the corrosion attack will occur. The area of
higher electrical potential is the cathode, or the positive pole, and is normally
not subject to corrosive attack. A difference of electrical potential will result
from two different metals or alloys or it may occur between two points on
the same material, due to local defects in chemical composition, a variation
of mechanical properties from stress or machining, or variations in the environment, such as a partially submerged component.
An electrolyte is a liquid that conducts electricity. Most liquids will conduct
electricity and therefore are good electrolytes. Seawater is an excellent electrolyte. Meanwhile, pure hydrocarbons are nonpolar, that is, they will not
conduct electricity. Therefore, metals immersed in hydrocarbons generally
are not subject to corrosion.
The metallic connection between the anode and the cathode provides a
path for the flow of electrons from the anode to the cathode and allows the
current to flow from the cathode to the anode. It can be a separate metallic
connection, or merely exist by the fact that the cathode and anode are in
contact with each other.
When corrosion occurs on a metal component that contains iron (Fe), the
result is the formation of Fe(OH)3 (ferric hydroxide), otherwise known as
rust. Meanwhile, on the cathode, a layer of H2 gas at the cathode surface
restricts corrosion. This corrosive-resistant layer of gas can be removed by
velocity or abrasion.
Although all corrosion on metals is electrochemical in nature as just
described, it is possible nevertheless to classify corrosion by type from the
visual appearance and the environment in which it takes place. These distinct types are caused by specific influences and specific environmental factors. The following list is generally accepted by most corrosion engineers to
include the common forms of corrosion in metals.
• Galvanic or two-metal corrosion
• Uniform or general corrosion
• Pitting corrosion
• Intergranular corrosion
• Erosion corrosion
• Stress corrosion
• Crevice corrosion
• Graphitization or dezincification corrosion
These corrosion types are discussed in more detail below.
© 2008 Taylor & Francis Group, LLC
Special Pump-Related Topics
217
A. Galvanic or Two-Metal Corrosion
When two metals separated by a spread in electrical potential are connected
in an electrolyte, the one with the lower potential becomes the anode and
will corrode. Meanwhile, the higher potential part, the cathode, will not
corrode. Table 7.1 shows the electrochemical order (also called the galvanic
series) of a number of common metals in a seawater environment. The higher
potential part is also called the more noble of the two materials. Note that
some materials, such as stainless steel and nickel, come in two forms, active
and passive, with widely varying electrical potential.
When there is a potential for galvanic corrosion in pump parts, the following practices should be observed to help restrict or reduce the amount of
corrosion.
• If two different metals will be wetted and in contact with one another
in a pump, select a combination of metals as close together as possible in the galvanic series. The corrosion rate is greater the farther
apart the two metals are in the galvanic series.
• Avoid the combination of a small anode in contact with a large cathode. The higher the ratio of cathode size to anode size, the more
accelerated the corrosion rate of the anode.
• Insulate the two materials, if possible, to eliminate the metallic connection. The insulation can be in the form of plastic washers, gaskets, or sleeves.
• Avoid wet threaded areas or other crevice areas with two dissimilar
materials.
• Be careful about using metal coatings that are high in the galvanic
series. A very small pinhole in the high galvanic metal coating will
result in a large cathode/small anode relationship. Materials lower
in the galvanic series, such as zinc, often make superior corrosionresistant coatings.
• Install a third (sacrificial) metal that is anodic to both working metals. This form of cathode protection is commonly used to protect
large metal tanks and other metal structures. Aluminum, zinc, and
magnesium are commonly used as the anode material.
B. Uniform or General Corrosion
This is a type of corrosion that occurs uniformly over the entire exposed metal
surface. The metal becomes thinner as the corrosion works its way through the
part and eventually the part fails mechanically. A plate immersed in an acid
will dissolve at a uniform but rapid rate. A steel plate exposed to the atmosphere will rust at a uniform, although much slower rate. Uniform corrosion
is by far the most common corrosion attack in metals. Because the metal loss
© 2008 Taylor & Francis Group, LLC
218
Pump Characteristics and Applications
TABLE 7.1
Galvanic Series of Common Metals and Alloys
in Seawater
Most noble or cathodic
Least noble or anodic
Platinum
Gold
Graphite
Titanium
Silver
Hastelloy C
Stainless steel (passive)
Nickel (passive)
Monel
Bronze
Copper
Brass
Hastelloy B
Nickel (active)
Tin
Lead
Stainless steel (active)
Ni-Resist
Cast iron
Steel
Aluminum
Cadmium
Zinc
Magnesium
is uniform based on the environment, it is the most predictable type for common applications. Corrosion handbooks provide data on suitability of different
metals for different corrosive environments, along with corrosion loss in mils
per year for different metals exposed to different corrosive liquids. The corrosion loss (in mils per year) is multiplied by the design life (in years) to obtain
the corrosion allowance. The resulting corrosion allowance is then added to the
thickness of the material that is required for structural or functional purposes,
and the result is the minimum thickness that the part can be.
C. Pitting Corrosion
Pitting corrosion is a local spot where the surface protection is attacked at
an isolated location. It is the prevalent form of attack with passive metals,
but it can occur on any metal under the right conditions. Pitting corrosion
involves the localized breakdown of the passive film that otherwise protects
the metal. This area is adjacent to a large area of high potential or cathode,
© 2008 Taylor & Francis Group, LLC
Special Pump-Related Topics
219
with the result that an accelerated corrosive attack occurs in the area of the
pit. Pitting generally occurs in an environment where the particular alloy
ordinarily exhibits only a moderate corrosion rate. Frequently, the breakdown in a passive film is caused by a minor trace element, such as fluorine,
chlorine, bromine, or iodine, with chlorine being the most common.
D. Intergranular Corrosion
Intergranular corrosion is a localized attack, occurring at or near grain
boundaries, with very little corrosion of the base material. The austenitic
(300 series) stainless steels are particularly susceptible to this type of corrosive attack, especially if they are not properly heat treated. If the material
is allowed to remain too long in the temperature range between 950°F and
1450°F (510°C and 788°C), known as the heat sensitization range, a chromium
carbide precipitate forms along the grain boundaries. At the same time, this
causes areas adjacent to the grain boundaries to become depleted of chromium. The chromium-depleted area next to the grain boundary becomes
an anode, adjacent to a large chromium-rich cathode, and the anode area is
subject to rapid localized corrosion in environments in which stainless steel
would not normally be expected to corrode.
Intergranular corrosion can be prevented by proper heat-treating (solution
annealing) of the material. The heat treatment consists of heating the part to
between 1950°F and 2050°F (1066°C and 1121°C), followed by a quick quench
to ensure transition through the heat sensitization range in less than 3 min.
In cases where the part is too big or not practical to heat treat, another preventative method is to lower the carbon content to 0.03% maximum (such as
the case with 304- or 316-L stainless steel). Still another way to prevent this
corrosion attack is by adding stabilizing elements, such as columbium (347
stainless) or titanium (321 stainless).
E. Erosion Corrosion
Erosion corrosion is the wearing away, by high velocity or abrasion, of the
protective surface film that is resistant to corrosion. Thus, this attack takes
place with a combination of corrosion and erosion. Metal loss associated
with cavitation is a form of erosion corrosion. With an iron or steel impeller
subject to cavitation, a corrosive film of Fe(OH)3 (rust) forms on the surface
of the impeller vane. Along comes a bubble and, when it collapses, the collapsing bubble causes destruction of the protective film and removal of the
unprotected material. Then a new rust film forms, and the process keeps
repeating until ultimately the damaged area gets large enough to be critical.
Another example of erosion corrosion is the effect of velocity on the corrosion rate of a particular part exposed to moving liquid. For some materials,
there is a breakaway velocity above which the corrosion rate rapidly accelerates. Other materials (e.g., titanium) show almost no change in corrosion
© 2008 Taylor & Francis Group, LLC
220
Pump Characteristics and Applications
rate even when exposed to high fluid velocities. Still other materials actually
exhibit a higher acceleration rate at lower velocities than at higher velocities
(e.g., stainless steel in slow-moving or stagnant seawater will corrode faster
than in higher-velocity seawater).
F. Stress Corrosion
Stress corrosion occurs when a metal is subjected to a combination of
mechanical stress and a corrosive medium. The condition is accelerated
at higher temperatures. Chemical environments that are likely to produce
stress corrosion cracking include chlorides, caustics, and ammonia. The
stress in the metal component can be induced by a combination of pressure,
heat treatment, machining, and/or forming.
G. Crevice Corrosion
Crevice corrosion occurs at a crevice in a pump part (typically occurring at
a place where two surfaces that are the same material are in loose contact
or are separated by a gasket that has allowed wetting of the gasket surface).
The crevice typically has an area of depleted oxygen, while the liquid being
pumped past the crevice is richer in oxygen. Therefore, there is localized corrosion in the portion of the crevice that is depleted of oxygen.
H. Graphitization or Dezincification Corrosion
Graphitization corrosion occurs when one element of a material is selectively removed, leaving a part that has more or less the same appearance
as the original part, but is porous and has considerably reduced physical
properties. This corrosion type occurs when cast iron is exposed to certain
acid water environments. In this environment, iron is selectively removed,
leaving a part that has a higher percentage of carbon (graphite) than it previously did, although the part may not look any different. Because of the
TABLE 7.2
Recommended Pump Materials for Different
pH Liquids
pH Value
10–14
8–10
6–8
4–6
0–4
Materials of Construction
Corrosion-resistant alloys or nonmetals
All iron
Bronze fitted or standard fitted
All bronze
Corrosion-resistant alloys or nonmetals
© 2008 Taylor & Francis Group, LLC
Special Pump-Related Topics
221
higher graphite composition, the part may fail mechanically because it is
not as strong.
Dezincification, a similar corrosion process, can occur in the high zinc
bronzes, where selective loss of zinc can occur.
Material selection on pumps is often performed by an expert on metallurgy and corrosion. In the absence of such expertise, Table 7.2 outlines some
generalized recommended materials for pumps handling liquid of known
pH value.
V. Nonmetallic Pumps
Industrial equipment made of plastic used to be thought of as a cheap but
weaker alternative, resulting in shorter service life. As a result, nonmetallic components were seldom specified by industrial pump users. The
past 20 to 25 years have witnessed the introduction of many new plastic materials that are uniquely suited for nonmetallic pump components
and pumps made entirely of plastic. With the proper material selection,
the use of nonmetallic components can result in several benefits, which
include superior corrosion and abrasion resistance, extended service life,
elimination of contamination in ultrapure applications, lower weight, and
reduced cost.
The selection of the correct plastic for a particular application requires
careful attention to all of the application parameters, as is the case with the
selection of materials in a metal pump. The most important factors to consider when selecting plastic pumps or components are resistance against corrosion and abrasion of the particular plastic in the specific liquid, the liquid
temperature and pressure range to which the material will be exposed, the
liquid velocity, and the variation of stresses to which the components will be
exposed.
Plastics are broadly categorized as either thermoplastics or thermosets,
depending on the nature of their molecular structure. Thermoplastics offer
greater resistance to corrosion and abrasion and can be used in ultrapure
applications (such as ultrapure water used for computer chip manufacturing
or food-grade applications) in their virgin form. Thermosets, while having
lower resistance to abrasion and often to corrosion, have higher mechanical
tensile strength.
Table 7.3 summarizes the most popular plastics currently available for use
in many pump types, including some comments on their major application
strengths, weaknesses, and limitations.
In addition to the plastics cited in Table 7.3, other nonmetallic materials
can be used in pumps to extend service life. For example, ceramics such as
© 2008 Taylor & Francis Group, LLC
222
Pump Characteristics and Applications
TABLE 7.3
Application Highlights for Plastic Pump Components
Material
PVC (polyvinyl
chloride)
Chemical Resistance
Upper
Temperature
Limit (°F)
Resists many acids,
alkalis, and other
chemicals
Similar to PVC
140
PP (polypropylene)
Useful in many acid,
alkali, and solvent
services
185
PE (polyethylene)
Similar to PP
200
PVDF
(polyvinylidene
fluoride)
Useful for most acids,
alkalis, solvents, and
many halogens
275
ECTFE (ethylene
chlorotrifluoroethylene)
PTFE
(polytetrafluoroethylene),
commonly called
Teflon®
FRP/GRP
(fiberglass or
glass-reinforced
polyester)
Similar to but better
than PVDF
300
Broadest chemical
resistance commonly
available for pumps
400
Useful in many
corrosive services
230
CPVC (chlorinated
polyvinyl chloride)
212
Other Important
Characteristics
Relatively low cost
Not useful against many
solvents
Superior in abrasion
resistance and stronger
than PVC
Relatively low cost
Poor with strong
oxidizing acid or
chlorinated
hydrocarbons
Lightest of the
thermoplastics
Similar in mechanical
properties to PP, but not
as light in weight
Strong and abrasion
resistant
Excellent in virgin state
for ultrapure services
Beats PVDF in abrasion
resistance and for
ultrapure applications
Highest cost of
nonmetallic choices
Not very resistant to
abrasives
chrome oxide can be applied at shaft journals under sleeve bearings (such
as in vertical turbine pumps as described in Chapter 4, Section XI) or onto
shaft sleeves of any pump fitted with packing, to achieve superior abrasion
resistance compared with most metals. The ceramic material has very weak
mechanical properties but is subject to almost no mechanical loading when
used as a coating on a shaft or sleeve. The proper preparation of the surface
to be coated with the ceramic and the application of this coating is essential
to its good service life. Ceramic coatings can handle highly corrosive environments and can operate in temperatures above 500°F.
© 2008 Taylor & Francis Group, LLC
Special Pump-Related Topics
223
VI. Materials Used for O-Rings in Pumps
A. General
Before discussing materials most commonly used for O-rings in pumps, a
few terms should be defined. The following material on O-rings and O-ring
materials is printed with permission of the Parker Hannifin Corporation,
Seal Group.
1. Polymer
A polymer is the result of a chemical linking of molecules into a long chainlike structure. Both plastics and elastomers are classified as polymers. In
this book, polymer generally refers to a basic class of elastomer, members
of which have similar chemical and physical properties. O-rings are made
from many polymers, but a few polymers account for the majority of O-rings
produced, namely nitrile (Buna N), EPDM, and neoprene.
2. Rubber
Rubber-like materials first produced from sources other than rubber trees were
referred to as “synthetic rubber.” This distinguished them from natural gum
rubber. Since then, usage in the industry has broadened the meaning of the
term “rubber” to include both natural as well as synthetic materials having
rubber-like qualities. This book uses the broader meaning of the word “rubber.”
3. Elastomer
Although elastomer is synonymous with rubber, it is formally defined as a
“high molecular weight polymer that can be, or has been modified, to a state
exhibiting little plastic flow and rapid, and nearly complete recovery from
an extending or compressing force.” In most instances, we call such material
before modification “uncured” or “unprocessed” rubber or polymer.
When the basic high molecular weight polymer, without the addition of plasticizers or other dilutents, is converted by appropriate means to an essentially
nonplastic state and tested at room temperature, it usually meets the following
requirements to be called an elastomer. The American Society for Testing and
Materials (ASTM) uses these criteria to define the term “elastomer.”
• It must not break when stretched approximately 100%.
• After being held for 5 min at 100% stretch, it must retract to within
10% of its original length within 5 min of release. (Note: Extremely
high hardness/modulus materials generally do not exhibit these
properties even though they are still considered elastomers.)
© 2008 Taylor & Francis Group, LLC
224
Pump Characteristics and Applications
4. Compound
A compound is a mixture of base polymer and other chemicals that form
a finished rubber material. More precisely, a compound refers to a specific
blend of chemical ingredients tailored for particular required characteristics
to optimize performance in some specific service.
The basis of compound development is the selection of the polymer type.
There may be a dozen or more different ones from which to choose. The rubber compounder may then add various reinforcing agents such as carbon
black, curing, or vulcanizing agents such as sulfur or peroxide, activators,
platicizers, accelerators, antioxidants, or antiozonants to the elastomer mixture to tailor it into a seal compound with its own distinct physical properties. Because compounders have thousands of compounding ingredients at
their disposal, it seems reasonable to visualize 2, 3, or even 100+ compounds
having the same base elastomer, yet exhibiting marked performance differences in the O-ring seal.
The terms compound and elastomer are often used interchangeably in a
more general sense. This usage usually references a particular type or class
of materials such as nitrile compounds or butyl elastomers. Please remember
that when one specific compound is under discussion in this book, it is a
blend of various compounding ingredients (including one or more base elastomers), with its own individual characteristics and identification in the form
of a unique compound number.
B. Eight Basic O-Ring Elastomers
The following are brief descriptions of eight of the most commonly used
O-ring elastomers. There are, of course, many other specialized polymers.
The ones listed below, however, account for well over 95% of all O-rings.
1. Nitrile (NBR, Buna N)
Due to its excellent resistance to petroleum products, and its ability to be
compounded for service over a temperature range of −65°F to +275°F (–54°C
to +135°C), nitrile is the most widely used elastomer in the seal industry
today. Most military rubber specifications for fuel- and oil-resistant MS and
AN O-rings require nitrile-based compounds. It should be mentioned, however, that to obtain good resistance to low temperature in a nitrile material,
it is almost always necessary to sacrifice some high-temperature fuel and oil
resistance. Nitrile compounds are superior to most elastomers with regard to
compression set, cold flow, tear, and abrasion resistance. Inherently, nitrilebased compounds do not possess good resistance to ozone, sunlight, or
weather. However, this specific weakness has been substantially improved
through compounding efforts.
© 2008 Taylor & Francis Group, LLC
Special Pump-Related Topics
225
Nitrile is recommended for general purpose sealing, petroleum oils and
fluids, water, silicone greases and oils, di-ester based lubricants (MIL-L7808), and ethylene glycol-based fluids (hydrolubes).
2. Neoprene
Neoprene can be compounded for service at temperatures from −65°F to
+250°F (–54°C to +121°C). Most elastomers are either resistant to deterioration from exposure to petroleum lubricants or oxygen. Neoprene is unusual
in having limited resistance to both. This characteristic, combined with a
broad temperature range and moderate cost, accounts for its use in many
sealing applications.
Neoprene is recommended for refrigerants (freons, ammonia), high aniline point petroleum oils, mild acid resistance, and silicate ester lubricants.
3. Ethylene Propylene (EP, EPR, and EPDM)
Ethylene propylene has won broad acceptance in the sealing world because
of its excellent resistance to Skydrol and other phosphate ester type hydraulic fluids. Ethylene propylene has a temperature range from −65°F to +300°F
(−54°C to +150°C) for most applications.
Ethylene propylene is recommended for phosphate ester based hydraulic
fluids (Skydrol, Fyrquel, Pydraul), steam to +400°F (+204°C), water, silicone
oils and greases, dilute acids, dilute alkalis, ketones (MEK, acetone), alcohols,
and automatic brake fluids.
4. Fluorocarbon (FKM, Viton, and Kalrez)
Fluorocarbon elastomers were first introduced in the mid-1950s. Since that
time, they have grown to major importance in the seal industry. Due to their
wide-spectrum chemical compatibility and broad temperature range, fluorocarbon elastomers represent one of the most significant elastomer developments in recent history.
The working temperature range of fluorocarbons is considered to be from
−20°F to +400°F (−29°C to +204°C), but some formulations have been known
to seal at −65°F (−54°C) in some static low-temperature applications.
Recent developments in material formulation have further improved
the characteristics of this very useful seal material. Fluorocarbon materials should be considered for use in aircraft, automotive, and other devices
requiring maximum resistance to deterioration by environment and fluids.
Fluorocarbon is recommended for petroleum oils, di-ester based lubricants (MIL-L-7808, MIL-L-6085), silicate ester based lubricants, silicone fluids
and greases, halogenated hydrocarbons, selected phosphate ester fluids, and
acids.
© 2008 Taylor & Francis Group, LLC
226
Pump Characteristics and Applications
5. Butyl
Prior to the introduction of ethylene propylene, butyl was the only elastomer
that was satisfactory for Skydrol 500 service over a temperature range from
−65°F to +225°F (−54°C to +107°C).
In addition, butyl exhibits excellent resistance to gas permeation, which
makes it particularly useful for vacuum applications.
Butyl is recommended for phosphate ester type hydraulic fluids (Skydrol,
Fryquel, Pydraul), ketones (MEK, acetone), silicone fluids and greases, and
for vacuum service.
6. Polyacrylate
This material has outstanding resistance to petroleum-based fuels and oils.
In addition, polyacrylate has good resistance to oxidation, ozone, and sunlight, combined with an excellent ability to resist flex cracking. Compounds
of polyacrylate have been developed that are suitable for continuous service
in hot oil over a temperature range from 0°F to +300°F (–18°C to +150°C).
Resistance to hot air is slightly superior to nitrile polymers, but tear
strength, compression set, and water resistance are inferior to many other
polymers. There are several polyacrylate types available commercially, but
all are essentially polymerization products of acrylic acid esters.
The greatest use of polyacrylate elastomers is by the automotive industry
in automatic transmission and power-steering devices using Type A transmission fluids.
Polyacrylate is recommended for petroleum oils, automatic transmission
fluids, and power-steering fluids.
7. Silicone
The silicones are a group of elastomeric materials made from silicon, oxygen, hydrogen, and carbon. As a group, the silicones have rather poor tensile strength, tear, and abrasion resistance. Special silicone compounds
have been developed that exhibit exceptional heat and compression set
resistance. High-strength materials have also been developed but their
strength does not compare with conventional elastomers. Silicones have
excellent resistance to temperature extremes. Flexibility below −175°F
(−114°C) has been demonstrated, and there are compounds that will resist
temperatures up to +700°F (+371°C) for short periods. The maximum temperature for which silicones are recommended for continuous service in
dry air is +450°F (+232°C). The ability of silicone to retain its original physical properties at these high temperatures is superior to most other elastomer materials.
Silicone compounds are not normally recommended for dynamic sealing
applications due to silicone’s rather low abrasion resistance.
© 2008 Taylor & Francis Group, LLC
© 2008 Taylor & Francis Group, LLC
Y
A
E
V
L
I
R
C
N
N
V
A
T
P
G
S
Epichlorohydrin
Ethylene acrylic
Ethylene propylene
Fluorocarbon
Fluorosilicone
Isoprene
Natural rubber
Neoprene
HNBR
Nitrile or Buna N
Perfluorinated fluoroelastomer
Polyacrylate
Polysulfide
Polyurethane
SBR or Buna S
Silicone
Abrasion
Resistance
P
G
E
P
G
P
G
G
G
E
E
P
G
GE
F
G
G
G
FG
E
GE
E
Acid
Resistance
FG
F
P
P
P
E
F
E
FG
FG
FG
FG
E
G
F
FG
G
F
G
FG
E
Chemical
Resistance
GE
FG
FG
G
P
E
FG
FG
FG
FG
FG
E
E
E
FG
G
E
FG
E
FG
P
E
G
G
G
P
PF
G
G
FG
G
G
GE
PF
GE
G
GE
FG
PF
G
G
P
G
E
F
F
F
GE
GE
F
E
F
P
GE
GE
F
G
F
G
F
F
G
Dynamic
Properties
Source:
Parker Hannifin Corporation, Seal Group, Lexington, KY. With permission.
Note: P = poor; F = fair; G = good; E = excellent.
K
H
Chlorosulfonated polyethylene
B
Butyl
Chlorinated polyethylene
V
D
Butadiene
Parker
Compound
Prefix Letter
AFLAS (TFE/Prop)
Elastomer Type (Polymer)
Cold
Resistance
Comparison of Properties of Commonly Used Elastomers
TABLE 7.4
E
Electrical
Properties
E
G
FG
F
F
E
F
F
F
G
G
E
F
G
F
F
F
G
G
G
Flame
Resistance
F
P
P
P
P
E
P
P
G
P
P
G
E
P
P
FG
G
GE
P
P
E
Heat
Resistance
E
FG
F
P
E
E
G
E
G
F
F
E
E
E
E
FG
G
G
G
F
E
Impermeability
P
F
G
E
E
G
G
G
G
F
F
P
G
G
E
GE
G
G
E
F
G
Oil Resistance
FG
P
G
E
E
E
E
E
FG
P
P
G
E
P
F
E
F
FG
P
P
E
Ozone
Resistance
E
P
E
E
E
E
P
G
GE
P
P
E
E
E
E
E
E
E
GE
P
E
Set Resistance
GE
G
F
P
F
G
GE
GE
F
G
G
GE
GE
GE
G
PF
F
F
FG
G
G
PF
Tear Resistance
P
FG
GE
P
FG
PF
FG
FG
FG
GE
GE
P
F
GE
F
G
G
FG
G
GE
Tensile
Strength
P
GE
E
F
F
FG
GE
E
G
E
E
F
GE
GE
G
G
F
G
G
E
FG
Water/Steam
Resistance
F
FG
P
F
P
GE
FG
E
F
FG
FG
F
FG
E
PF
F
F
F
G
FG
GE
Weather
Resistance
E
F
E
E
E
E
F
G
E
F
F
E
E
E
E
E
E
E
GE
F
E
Special Pump-Related Topics
227
228
Pump Characteristics and Applications
Silicones are recommended for high-aniline point oils, dry heat, and chlorinated diphenyls.
8. Fluorosilicone
Fluorosilicone elastomers combine the good high- and low-temperature
properties of silicone with basic fuel and oil resistance. The primary uses
of fluorosilicone are in fuel systems at temperatures up to +350°F (+177°C)
and in applications where the dry heat resistance of silicone is required but
the material may be exposed to petroleum oils and/or hydrocarbon fuels.
The high temperature limit for fluorosilicone is limited because temperatures approaching +350°F (+177°C) may degrade certain fluids, producing
acids that attack the fluorosilicone elastomer. High-strength fluorosilicone
materials are available and certain ones exhibit much improved resistance
to compression set.
Fluorosilicone is recommended for petroleum oils and fuels.
Styrenebutadiene rubber (SBR)
Polyurethane rubber (AU, EU)
Butyl rubber (IIR)
Low-temperature nitrile rubber (NBR)
Hydrogenated nitrile rubber (HNBR)
High-temperature nitrile rubber (NBR)
Chloroprene rubber (CR)
Polyacrylate rubber (ACM)
Ethylenepropylenedienerubber (EPDM)
Fluorosiliconerubber (FMQ, FVMQ)
TFE/Propropylene rubber (FEPM)
Fluorocarbon rubber (FKM)
Perfluorinated elastomer (FFKM)
Siliconerubber (VMQ)
°C –100 –75
°F –148 –103
–50
–58
–25
–13
0
32
25
77
50
122
75
100
167
212
Tempe rature
Normal recommended temperature range
125
257
150
302
175
347
200
392
225
437
250
482
300
572
Extended temperature range for short term only
FIGURE 7.1
Temperature range for common elastomeric materials. (Courtesy of Parker Hannifin
Corporation, Seal Group, Lexington, KY.)
© 2008 Taylor & Francis Group, LLC
Special Pump-Related Topics
229
Table 7.4 summarizes and compares the properties of commonly used elastomers found in O-rings, including those described above. Figure 7.1 summarizes the temperature ranges for common elastomeric materials.
VII. High-Speed Pumps
The high-speed centrifugal pump has roots in the aerospace industry and in
military applications. The pump (Figure 7.2) uses an integral gearbox drive
FIGURE 7.2
High-speed centrifugal pump. (Courtesy of Sundyne Corporation, Arvada, CO.)
© 2008 Taylor & Francis Group, LLC
230
Pump Characteristics and Applications
to achieve speeds up to 25,000 rpm. This allows the pump to achieve heads to
about 6300 ft with a single stage, with flows up to about 400 gpm. Flow rates
up to 1000 gpm are achievable at lower heads with a single stage. Two pumps
in series driven from a common gearbox allow heads to about 15,000 ft at 400
gpm or to about 6300 ft at 1000 gpm.
Applications for high-speed, high-head pumps are principally in the refining and petrochemical industries. These pumps are also used in paper processing, power generation, steel production, and other heavy-duty industrial
applications.
Many high-speed pump designs use what is called a partial emission design
(Figure 7.3). This design uses impellers whose vanes are oriented in a straight
radial direction, called Barske impellers, rather than being curved backward
like most impellers. Also, the partial emission design has the impeller concentrically located in the casing, and has a diffuser that only allows flow
from a small portion of the impeller to be delivered to the pump discharge at
any one time. Partial emission designs are able to produce higher efficiency
pumps at low specific speeds. Beyond the range of partial emission impellers, full emission impellers using traditional impeller designs with double
volutes or diffusers are employed.
High-speed centrifugal pumps have a number of advantages compared
with the alternatives of single-stage centrifugal pumps with larger impeller
diameters, multistage centrifugals, multiple pumps operating in series, or
reciprocating positive displacement pumps. The advantages include higher
efficiency, reduced sensitivity to dry running due to the larger running clear-
FIGURE 7.3
Partial emission pumps can achieve higher efficiencies at low specific speeds. (Courtesy of
Sundyne Corporation, Arvada, CO.)
© 2008 Taylor & Francis Group, LLC
Special Pump-Related Topics
231
ances, and reduced size and weight. The smaller size and weight can mean
lower first cost than other alternatives, particularly in alloy construction.
There are several areas that deserve special attention with this unique
pump design. These include the fact that the gearboxes required to produce the high speed necessitate additional cost and maintenance for seals,
bearings, and gear sets. Some VFD designs are available to achieve higher
speeds without the use of gearboxes. Also, high-speed designs may require
inducers to achieve a reasonable NPSHr. The dynamic balance of the rotor
and the alignment among the pump, gearbox, and motor are much more
critical at the high speeds at which these pumps operate. Finally, selection
of the proper mechanical seal is more critical and service is more severe
because the seals are running at higher surface speeds. This last area of
concern can be eliminated as sealless options become more available for
high-speed pumps.
VIII. Bearings and Bearing Lubrication
A number of advances have been made in the field of bearing design and
bearing lubrication systems through the years. More precise manufacturing techniques have permitted the life of traditional ball- and roller-type
bearings to be significantly increased. Plate-type thrust bearings can be used
in pump types that generate higher thrust loads than are capable of being
handled with traditional ball- and roller-type bearings. Precision alignment
techniques, discussed in Section IX below, have added considerably to the
expected life of many bearing systems.
The newest development in the area of bearing design is in the area of
magnetic bearings. The use of a magnetic field to center a pump shaft and to
carry the radial and thrust loads of a pump offers several major benefits. The
lack of mechanical contact in the bearing system increases pump efficiency,
extends bearing life, reduces maintenance, and eliminates the need for bearing lubricants. For sealless centrifugal pumps (magnetic drive and canned
motor pumps), magnetic bearings may also allow dry running, pumping
liquid containing abrasives, as well as operation over a wider flow range,
because the bearings do not depend on the pumped liquid for lubrication.
The technology of magnetic bearings is still in its infancy, with few working prototypes in the field. The prospects are exciting, however, and may
be the breakthrough that is necessary to allow magnetic drive and canned
motor technologies to reach their full potential.
In the area of bearing lubrication, oil mist lubrication of bearings for a wide
range of equipment in large plants is being used with more and more regularity. This lubrication system, often based on a central system distributing
an oil mist to a variety of pieces of rotating equipment in the plant, offers a
© 2008 Taylor & Francis Group, LLC
232
Pump Characteristics and Applications
number of potential benefits compared with traditional grease or oil lubrication of individual pump bearings. These benefits can include significantly
longer bearing life, a reduction in the amount of bearing oil consumed, and
a reduction in oily waste pollution.
IX. Precision Alignment Techniques
The importance of coupling alignment for centrifugal pumps is discussed
in Chapter 8. Lack of proper alignment may very well be the single most
important cause of premature failure of frame-mounted pumps. Like many
other activities, the degree of pump alignment is not a specific point, but
rather a spectrum, with more precise alignment generally being reserved
for larger, higher-speed, more expensive, and more critical equipment. To
achieve the full range of possible alignment accuracies, many approaches
can be used, ranging from very simple and rudimentary alignment using
straight edges and feeler gauges to the use of laser alignment equipment at
the other end of the spectrum. In between these two extremes of the spectrum are other alignment techniques, such as single dial indicator (rim and
face) alignment, reverse indicator alignment, and alignment using electronic
gauges. The details of these alignment methods and instruments are beyond
the scope of this book, but the interested reader is referred to Ref. [10] for
more information on alignment techniques and instrumentation.
Two of the terms used to express the degree of misalignment are offset
and angle. Offset indicates the difference between the two shaft centerlines
at the coupling center or the amount of parallel misalignment, expressed in
mils (1 mil = 0.001 in.). Angle is the change in gap between the coupling
faces, divided by the distance across the faces, or the angular misalignment,
expressed in mils per inch or degrees. Angular and parallel misalignment
are illustrated in Chapter 8, Figures 8.1 and 8.2.
In recent years, there has been a significant increase in the use of optical
and laser alignment instruments. These devices have become more widely
available, more affordable, and easier to use. Added to that is the growing
understanding of the extension of the mean time between failure that can be
achieved by more precise alignment efforts. The benefits of these precision
alignment techniques include longer bearing life, longer seal life, improved
pump reliability, reduced maintenance costs, and lower noise levels.
The most precise and sophisticated alignment method is not always the
correct choice of alignment technique for every pump application. The decision of the degree of alignment accuracy is a tradeoff among alignment
precision (and presumably pump life), alignment time, and the cost of the
instrumentation used for alignment. Using the most sophisticated alignment techniques, misalignment can be measured far more accurately than is
© 2008 Taylor & Francis Group, LLC
233
Special Pump-Related Topics
TABLE 7.5
Pump Alignment Tolerances
rpm
Offset (mils)
Angle (mils/in)
600
900
1200
1800
3600
7200
5.0
3.0
2.5
2.0
1.0
0.5
1.0
0.7
0.5
0.3
0.2
0.1
practical to correct. Also, some of the most sophisticated alignment methods
require that the pump equipment be brought into a certain level of alignment, using simpler equipment and techniques, before the more precise
instrumentation can even be used.
Pump speed is also an indicator of the degree of alignment precision
required. For example, a pump running at 900 rpm requires much less accurate alignment than one running at 3600 rpm. If no other guidelines are
available from the pump manufacturer or from company policies, the alignment tolerances shown in Table 7.5 can be used.
X. Software to Size Pumps and Systems
Chapter 3, Section III, describes computer technology that allows the modeling and optimization of complex piping networks, so that pump head and
the reaction of the pump and system to various operating modes can be
accurately determined. These computer software programs make it much
more feasible to accurately analyze new piping systems during the design
phase, or to model existing pumping systems for modification or improvement. System designers and engineers will find the use of software particularly helpful and cost effective if the system is complex, has multiple flow
branches, employs multiple pumps, or contains liquids that are appreciably
different from water in terms of their effect on pipe friction losses and pump
performance.
© 2008 Taylor & Francis Group, LLC
8
Installation, Operation, Maintenance,
and Repair
I. Overview
This chapter provides a step-by-step procedure for installing and starting
up a centrifugal pump in the field. The most important consideration for
the successful installation and startup of many pumps is careful attention
to alignment. After the pump is started, the benchmarks used to monitor
performance throughout the life of the pump are established. These benchmarks include hydraulic performance, temperature, and vibration.
This chapter discusses the criteria that should be considered in determining minimum continuous flow rate of a pump. The criteria that might affect
the determination of minimum flow include temperature rise, radial and
axial thrust bearing loads, prerotation, recirculation, separation, settling of
solids, noise, vibration, and power consumption.
This chapter also includes a discussion of ten ways to prevent low flow
damage in pumps.
Regular and preventive maintenance techniques for pumps are discussed,
along with the benefits of establishing benchmarks of performance for new
or recently overhauled pumps.
If a problem does occur with an operating pump, this chapter outlines
some of the things that can be done to find the cause of the problem. A troubleshooting chart is provided as a guide for such an investigation.
Finally, if repair of the pump is necessary, this chapter provides some
guidelines for successful pump repair.
II. Installation, Alignment, and Startup
A. General
This section describes a step-by-step checklist for installing and starting up
a frame-mounted centrifugal pump. The order of the steps and the specific
235
236
Pump Characteristics and Applications
startup activities vary for each type of pump, and the reader is urged to consult the manufacturer’s instruction manual for the specific piece of equipment. For low-cost, light-duty pumps, it may not be cost effective to carry out
the multiple alignment checks and some of the other procedures discussed
in this section. Within the bounds of reason, however, attention to the spirit
of this checklist, if not to the letter, is recommended for all industrial pumping equipment.
For a pump supplied with a coupling, alignment is undoubtedly the most
important aspect of the startup procedure, and it is repeated a number of
times during the startup procedure, with each successive alignment check
being more precise than the preceding one. Further discussion on the subject of alignment can be found in Chapter 7, Section IX. Some users have the
mistaken impression that they can specify that a frame-mounted pump be
aligned at the factory, thus eliminating the need for careful alignment in the
field. This is not generally true because the vibration to which the equipment
is exposed during its shipment to the job site most likely causes the pump
to be knocked out of alignment. Also, the pump alignment can be further
compromised during activities such as the completion of the piping. There is
no substitute for careful field alignment.
Another aspect related to alignment that causes some confusion to certain
pump owners is the so-called flexible couplings with which some pumps are
equipped. Some users mistakenly believe that flexible pump couplings are
designed to withstand large amounts of misalignment, but this is not true.
Flexible couplings are meant to withstand only very minor amounts of misalignment (refer to Chapter 4, Section XV), and they should by no means be
considered a substitute for good coupling alignment.
B. Installation Checklist
1. Tag and Lock Out
This important safety function should always be the first step performed
before beginning work on any pump. The equipment should be tagged and
locked out at the starter or switch gear, so that it cannot possibly be started,
either remotely or locally, while the pump is being serviced.
2. Check Impeller Setting
The impeller may have been set at the factory, but it is a good idea to recheck
the setting prior to installing the pump. As pointed out in Chapter 4,
Section II.A, the axial setting of the impeller is most critical for open impellers, as this affects pump efficiency. A common axial setting for open impellers is 0.012 to 0.015 in. Also, Chapter 4, Section II.D, discusses the trade-off
between efficiency, on the one hand, and thrust load and stuffing box pressure, on the other, in setting of open impellers.
Installation, Operation, Maintenance, and Repair
237
For closed impellers, the axial setting is much less critical, with the setting
being merely to ensure that the impeller is not rubbing against the front or
back side of the casing. Many closed impellers have no method of adjusting
impeller setting. If there is a method for setting the impeller, it can be set 1/8
to 3/8 in. (depending on pump size) from the front side of the casing.
3. Install Packing or Seal
The mechanical seal can be installed at the factory, but this runs the risk of
damaging the faces due to the vibration the pump undergoes in shipment.
This is especially true for carbon seal faces (a very common seal face material), which are subject to cracking if they receive a shock in transit.
The packing should be changed as part of the pump startup procedure,
especially if the pump has been stored for an extended period before startup.
4. Mount Bedplate, Pump, and Motor
If the pump is to be mounted on a finished concrete surface, the surface
should be roughed up with a chipper, so that when grout is installed under
the bedplate (Step 6) it will adhere better to the concrete surface.
The pump bedplate should be mounted reasonably level on its foundation, using shims if required. Some bedplates are equipped with leveling
screws. The pump and motor are then mounted on the bedplate if this has
not already been done.
Note that for pumps smaller than about 40 HP, a stilt-mounted base can be
considered. This type of base requires no grouting, is above the floor so that
corrosion damage is minimal, and is designed to move as a unit with pipe
load, so that the pump, base, and motor maintain original alignment. This
last benefit has been questioned by some, whose experience has been that
the stilt-mounted base causes more pipe strain and deflection than a fully
grouted bedplate and ultimately shortens the life of mechanical seals.
Check the pump and motor for soft foot, a condition in which one of the
feet does not sit flat on the base (it may be parallel or angled). A dial indicator is positioned with the pin on the foot, then zeroed. The bolt at that foot is
loosened, and if the indicator moves away from zero, then shims are placed
under that foot. This procedure is repeated for all feet.
5. Check Rough Alignment
This first alignment check is done to ensure that the pump and motor can be
brought into more precise alignment later on with more sophisticated alignment methods. (See discussion on alignment techniques in Chapter 7, Section
IX.) Accordingly, this first rough alignment check can be performed with feeler
gauges between coupling halves to check angular alignment (Figure 8.1) and
a straight edge along coupling edges to check parallel alignment (Figure 8.2).
238
Pump Characteristics and Applications
Motor
Motor
Pump
Pump
FIGURE 8.1
Rough angular alignment check using feeler gauges.
Motor
Pump
Motor
Pump
FIGURE 8.2
Rough parallel alignment check using straight edge.
To bring the pump into parallel alignment from top to bottom, shims are
added or removed from under the motor feet. To adjust the side-to-side parallel alignment, the motor is shifted from one side of the bedplate to the other
(some larger setups have jacking screws to assist in this activity) as necessary
to bring it into alignment. For angular alignment, again shims are added or
removed as necessary from the rear or front feet of the motor, and the rear of
the motor is shifted from side to side as necessary.
Installation, Operation, Maintenance, and Repair
239
6. Place Grout in Bedplate
Most bedplates, whether they be cast or fabricated steel, have hollow cavities on their undersides. These cavities should be filled with grout to keep
the bedplate from flexing while the pump is operating and to ensure that all
loads are evenly transmitted to the foundation below the pump.
Grout should be allowed to fully dry, which generally takes 24 to 48 h,
before proceeding to the next step, because the bedplate may shift or flex
while the grout is drying. Check the bedplate for hollow cavities after the
grout has dried by tapping with a hammer. Fill in any remaining cavities
with additional grout (this may require drilling small holes in bedplate).
Stilt-mounted baseplates require no grouting, as discussed in Step 4.
7. Check Alignment
This alignment check should generally be done with a dial indicator, using
the rim and face method shown in Figure 8.3, or with a laser alignment tool.
(More precise alignment techniques are discussed in Chapter 7, Section IX.)
If the pump can operate at high temperature (or if the driver is a steam turbine that operates at a high temperature), the expected growth of the pump
or turbine due to thermal expansion must be estimated or obtained from the
supplier and taken into account in the parallel alignment procedure.
8. Flush System Piping
The system piping has not yet been connected to the pump. Before connecting the piping, it should be flushed to remove dirt, cutting chips, weld
FIGURE 8.3
Rim and face alignment technique using dial indicator.
240
Pump Characteristics and Applications
debris, or other foreign material from the piping. Temporary strainers can be
installed in the piping to collect this debris.
9. Connect Piping to Pump
If the system piping is brought up to the pump, there is a high probability that it will not be exactly in line with the pump suction and discharge
flanges (Figure 8.4). Applying leverage on the pipe to bring it in line with
the flanges is liable to cause undue strain on the pump flanges, which could
cause the stuffing box to deflect, eventually causing premature seal failure.
This can be avoided by bringing the system piping only up to the closest
pipe restraint rather than all the way up to the pump. Then, the piping is run
from the pump out past the closest pipe restraint. In this way, any strain that
has to be put on the pipe to get the last two flanges to connect is carried by
the pipe restraint, rather than by the pump flange.
When bolting the pump flange to the connecting piping, care should be
taken if the pump flange is a cast material like ductile iron or bronze. Because
the mating steel pipe often has a male register, this results in a gap between
the mating pipe and pump flanges. If the flange bolts are over-torqued, this
gap can cause the pump flange to break or to bend. Bolts should be tightened
snugly, but not overly, and should be tightened uniformly, a little at a time on
each bolt, rather than fully tightening a single bolt.
One way to avoid the possibility of breaking or bending the pump flange
is to machine off the register on the pipe flange that attaches to the pump
Wrong
Wrong
Right
FIGURE 8.4
If piping is not in alignment with pump flanges, this can lead to excessive deflection of the
pump stuffing box. (Courtesy of Goulds Pumps, Inc., a subsidiary of ITT Corporation.)
Installation, Operation, Maintenance, and Repair
241
flange, thereby eliminating the aforementioned gap. If the register on the pipe
is machined off, a full face gasket should be used between the two flanges.
The pressure-containing capability of a gasket is the bolt force used
divided by the gasket area. For the same bolt force, increasing the gasket
area (which would be done if the register is machined off the pipe flange
and a full face gasket is used) lowers the maximum pressure that the gasket can contain. This should be checked before the pipe flange register
is machined off to ensure that the resulting gasket pressure capability is
adequate.
Note that if the connecting pipe has a register on it, the gasket size is the
same whether or not the pump has a machined register. This means that
the pressure containing capability of the gasket is the same whether or not
the pump has a register. Specifying a raised face flange on a cast pump that
normally comes with a flat face flange does absolutely nothing in the way
of increasing the pressure carrying capability of the gasket, while probably
increasing the cost and delivery time for the pump.
10. Check Alignment
This alignment check is made to verify that no excessive strain has been
applied to the pump by the piping connection. If the alignment has changed
from the previous check (Step 7), the piping should be re-run.
11. Turn Pump by Hand
It should be possible to turn a new or newly overhauled pump by hand
(with perhaps the aid of a strap wrench for larger pumps). Note that vertical
pumps that have not had their impeller setting adjusted yet would not be
able to be turned by hand, since the impeller(s) would be resting on the casing immediately below it.
Turning the pump by hand, the technician makes note of any unusual
sounds or rubbing. This could be an indication of foreign matter such as dirt,
weld spatter, or cutting chips located in the pump; a bent shaft; or a pump
wear ring that is not machined concentrically or that has been installed
improperly. Any unusual observation should be thoroughly checked prior
to proceeding.
12. Wire and Jog Motor
Before the coupling halves between the pump and motor are connected,
the motor wiring is completed, and the motor jogged, or bump started to
confirm that the motor has the correct rotation. This is especially important
for pumps such as ANSI pumps that typically have the impeller threaded
onto the end of the shaft, tightening with rotation. If this type of pump is
allowed to rotate in the opposite direction from that which is intended, it
242
Pump Characteristics and Applications
simply backs the impeller off until it hits the casing with the front side of
the impeller, possibly damaging both the impeller and casing. Many other
pump designs can be damaged if operated in reverse rotation, so this rotation check is important for all pumps.
13. Connect Coupling
Coupling halves are connected, completing the assembly of the pump unit.
14. Check Shaft Runout
Checking the shaft runout with a dial indicator after the alignment has been
finalized and the coupling halves connected is a final assurance against
improper assembly, a bent shaft, or imprecise machining of any of the mating components in the pump bearing housing, shaft, sleeve, or coupling. The
dial indicator should be mounted on the pump base and the runout reading
taken with the pin of the indicator located on the shaft at the pump seal, and
the shaft rotated by hand from the coupling end. A well-made pump should
be able to achieve 0.002 in maximum total indicated runout in one revolution
of the shaft. A runout in excess of this may not be permissible (for example,
with pumps built to the API 610 specification described in Chapter 4, Section
XIV.C). A higher runout can be tolerated unless it is otherwise specified, but
0.002 in is a worthy goal.
15. Check Valve and Vent Positions
The valves in the suction line should be fully open, and drain lines on the
pump should be closed. Vent connections on the pump should be open if the
suction pressure is above atmospheric pressure; if not, provisions must be
made to prime the pump (Step 17). Pump discharge valve should be closed on
most radial and mixed flow pumps. (Higher horsepower units often require
the discharge valve to be opened about 10%. Consult the instruction manual
or manufacturer’s recommendation for the specific pump.) Finally, all valves
associated with the pump lubrication and cooling systems should be opened.
16. Check Lubrication/Cooling Systems
Lubrication systems include lubrication of the pump and driver. Refer to further discussion of lubrication systems in Section IV.A.1 to follow.
17. Prime Pump if Necessary
A centrifugal pump mounted above the source of suction requires priming
prior to startup. Priming fills the suction piping and pump casing, and vents
the air out of the pump.
Installation, Operation, Maintenance, and Repair
Steam air,
or water
supply
243
Priming
ejector
To
waste
FIGURE 8.5
Use an ejector (shown) or vacuum pump for priming. (From Karassik, I.J. et al., Pump Handbook,
4th Ed., McGraw-Hill, Inc., New York, 2008. With permission.)
If the pump has a foot valve in the suction piping, priming consists of
filling the suction line from an external source while venting the casing. If
there is no foot valve, the suction line must be filled by pulling a vacuum
at the top of the pump casing, using either an ejector (Figure 8.5) or a vacuum pump. The vacuum lifts the liquid up the suction line, priming the
pump.
18. Check Alignment
This alignment check is done to ensure that filling the pump during the
priming process has not caused the alignment to shift.
19. Check System Components Downstream
Make certain that all system piping, components, and the discharge vessel
are ready to receive flow and pressure.
20. Start and Run Pump
Run for at least 1 h, at design operating temperature if possible.
244
Pump Characteristics and Applications
21. Stop Pump and Check Alignment
This final alignment check should be done hot (i.e., while still at operating
temperature) if possible, to make certain that the thermal expansion allowance was correctly made. As pointed out in Chapter 4, Section XIV.C on API
610, a centerline-mounted case design minimizes movement of the casing
due to thermal growth.
22. Drill and Dowel Pump to Base
If the pump operation is satisfactory and if the pump is within specification
on alignment and vibration, secure the pump and motor to the bedplate with
tapered dowel pins to help maintain the alignment.
23. Run Benchmark Tests
Now that the pump has been successfully started and while the pump is
optimally aligned and balanced, this is the best time to run the benchmark
tests. Benchmark tests are designed to create a signature of the performance
characteristics of the pump while all components are new (or newly rebuilt).
These performance characteristics can then be periodically remeasured and
compared with the benchmark characteristics as a way to monitor any deterioration of the pump.
The most commonly measured benchmarks are the pump’s hydraulic performance, temperature, and vibration. These are discussed in more detail in
Section IV.C below.
III. Operation
A. General
Probably the best advice that can be given relative to good pump operation
is that the operator should try to keep the pump operating at a healthy point
on the head–capacity curve. If the pump is allowed to run too far out on the
curve, it can experience problems with cavitation, cause excessive power load
on the motor, can cause the pump radial bearings to be excessively loaded,
and can have excessive noise and vibration. Refer to Section III.B below for
a discussion of the detrimental effects of operating a pump for an extended
period below the minimum recommended flow for that pump.
Aside from running in a healthy operating zone, other operating practices should focus on minimizing the energy consumed. Keep the pump
near its best efficiency point if possible. Variable-speed pumps should not
be run at a higher speed than necessary to meet system requirements.
Installation, Operation, Maintenance, and Repair
245
Excessive throttling should be avoided if possible, as this wastes energy
and will likely increase maintenance problems for valves and pumps.
Pumps should never be operated with a closed suction valve. Valves on the
discharge side should usually not remain closed for more than a minute,
although this varies with pump size. High-energy pumps larger than several hundred horsepower are generally not recommended to ever run with
the discharge valve completely closed but rather should be cracked open
about 10% as a minimum. Refer also to the discussion to follow on minimum continuous operating flow.
For shutdown of the pump, the pump should be de-energized prior to closing of any block valves. If the pump is pumping against a static head and if
there is no check valve in the pump discharge (as there should be), it may be
necessary to close the discharge block valve immediately prior to shutting
off the pump. There should be no delay between these two actions, however.
This means that if the controls are not local to the pump, this operation may
require two people.
Finally, operators of pumps should ensure that the equipment is well
maintained, which is the subject of Section IV to follow.
B. Minimum Flow
A determination of the minimum acceptable continuous operating flow of
a pump is necessary to set limits on control equipment and instrumentation, establish operational procedures, and determine the need for and size
of minimum flow bypass systems. Unfortunately, there is no single answer
to this question for every pump and every circumstance. Rather, each pump
must be carefully examined in the context of its application to arrive at the
minimum flow. For example, many small pumps handling ambient temperature water can safely operate at flows as low as 25% of best efficiency flow for
long periods. However, a much larger pump, or one handling a liquid near
its boiling point, must be operated at a much higher percent of best efficiency
flow, perhaps as high as 80% of BEP.
The factors that should be considered in arriving at the minimum continuous flow for a particular pump are discussed in the paragraphs that follow.
Not all of these considerations are applicable to every case. The determination of which of the factors discussed below dominates a particular pump
application is not easy, and good engineering judgment must be used. In
general, a conservative approach is to supply pump systems with a system
to prevent the pump from operating below its recommended minimum flow.
1. Temperature Rise
As liquid passes through a centrifugal pump, its temperature increases due
to several effects, including friction and the work of compression. Assuming
that all the heat generated remains in the liquid, the temperature rise ΔT
246
Pump Characteristics and Applications
increases as head is increased, and as flow and efficiency are decreased. At
normal operating points, this temperature rise is minimal, but it may start to
become significant at flows lower than about 15% of the BEP flow.
At very low flow rates, with pump efficiency being very low, temperature
rise is the highest. Temperature rise limits are determined by the difference
between NPSHa at the pump suction and the NPSH required by the pump.
An arbitrary but usually conservative practice calls for limiting temperature
rise in centrifugal pumps to about 15°F, although for pumps handling cold
liquids, a rise of 50°F or more may be quite acceptable. For light hydrocarbons with high vapor pressure and where NPSHa is close to NPSHr, it would
be wise to double-check to ensure that the temperature rise does not elevate
the liquid temperature to the boiling point.
2. Radial Bearing Loads
As discussed in Chapter 1, Section V, radial bearing loads can affect both
shaft deflection and bearing life. As Figure 1.7 shows, the radial loads are at
a minimum at the BEP and may be considerably higher at reduced flow rates.
This can limit minimum flow if the excessive radial bearing load causes
excessive shaft deflection and premature seal or bearing failure. Of course,
whether or not this is a factor depends on how conservatively the pump
bearing system was originally designed. As an example, many manufacturers of ANSI process pumps only offer two or three different bearing frames
for over 20 pump sizes. There may be six or seven pump sizes using the same
bearing frame. Obviously, the smallest pump size in a given group has a
greater design safety factor on its bearing system than does the largest pump
in the same group. Most modern pump designs allow flow in the range of
20% of the BEP flow without exceeding radial bearing load limits, but again,
this should be checked with the manufacturer.
3. Axial Thrust
Axial thrust is discussed in Chapter 4, Section II.D. Axial thrust at shutoff
is higher than the axial thrust at BEP, in the same proportion as the ratio of
pump head at shutoff to the head at BEP. As per the discussion above about
radial bearings, the design safety factor of the thrust bearing system must be
considered to determine if this factor might restrict minimum flow. Typical
modern pump designs allow flow down to roughly 20% of BEP flow without
significantly affecting thrust bearing life.
4. Prerotation
This is the change in the inlet velocity triangle as a pump operates off the
BEP, causing the liquid to spiral in the suction ahead of the impeller. The net
effect of pre-rotation is to lower the pump’s suction head and efficiency. The
Installation, Operation, Maintenance, and Repair
247
effect is more pronounced the farther to the left of BEP the pump operates
and is more pronounced on larger pumps.
5. Recirculation
Recirculation is a flow reversal at the suction and/or discharge tips of impeller vanes (Figure 8.6). Suction recirculation causes noise and cavitation-like
damage at the impeller inlet, while discharge recirculation can cause similar
damage on the pressure side of the outlet edge of the vane. The flow also
separates from the vanes at these points.
Other symptoms of recirculation include cavitation-like damage to the
shrouds near the vane outlets and unexpected shaft breakage or thrust bearing failure.
It is important to note that this cavitation-like damage occurs on the pressure side of the impeller vane. When looking at the inlet of an impeller, the
pressure side of the vane is the side that is usually not able to be seen directly
with the eye, but which requires a mirror to see behind the vane. This is the
opposite side of the impeller vane from that where conventional cavitation
damage occurs.
Suction and discharge recirculation damage can occur as the pump
moves away from its BEP. Although the onset of recirculation alone is not
necessarily a cause for concern, if the pump is allowed to operate too long
in a period where there is too much of this recirculation, it can cause damage to the pump. Thus, this is a criterion used for determining minimum
flow on some pump types. The pump types most likely to exhibit problems
FIGURE 8.6
Recirculation occurs at impeller inlet and/or outlet at capacities below BEP. (From Karassik, I.J.
et al., Pump Handbook, 4th Ed., McGraw-Hill, Inc., New York, 2008. With permission)
248
Pump Characteristics and Applications
with recirculation are high-energy pumps and pumps with high suction
specific speed. The definition of what constitutes a high energy pump is
not agreed upon by everyone in the pump industry. Many put the threshold level at around 100 HP per stage. Others use a higher figure, such as
200 HP per stage.
The percent of BEP capacity at which suction recirculation begins is a function of the pump specific speed Ns, the suction specific speed S, and whether
the pump is single suction, double suction, or multistage (Ref. [3]). At higher
values of both Ns and S, the minimum acceptable flow (as a percentage of
BEP) to prevent recirculation increases. Also, the flow at which recirculation
begins (as a percentage of BEP) is higher for double suction pumps than for
single suction pumps, and higher still for multistage pumps. For example,
for Ns = 1000 and S = 7500, the percent of BEP at which recirculation begins
is about 50%, 55%, and 65%, respectively, for single suction, double suction,
and multistage pumps. For the same value of Ns but S = 10,000, the percent of
BEP at which suction recirculation begins is about 63%, 73%, and 83% of BEP
for the same three pump types. For higher S values, these percentages are
even higher. This effect is one reason why large-eye impellers for reduced
NPSHr (higher S value) must be used with caution, as this can restrict the
safe operating range of the pump.
Note that there is neither complete agreement within the pump industry
as to exactly when the onset of recirculation begins for a given pump type,
nor is there complete agreement as to how the onset of recirculation should
relate to the recommended minimum flow for that particular pump. Some
suggest that the minimum continuous flow show not be less than one half of
the flow at which recirculation begins.
Some manufacturers are more conservative in their approach than others, and some industrial pump end users such as refinery owners tend to
be more conservative than some pump manufacturers in setting minimum
flows. A good communication between the pump end user and manufacturer is essential in establishing the recommended minimum flow to avoid
recirculation for a given pump, especially a high-energy pump.
6. Settling of Solids
For liquids containing solids, the minimum flow must be high enough to
prevent the solids in the liquid from settling in the pump or in the piping.
7. Noise and Vibration
At reduced flow rates, pump operation may be recirculation-induced noisy due
to a combination of the effects of prerotation, recirculation, and recirculationinduced cavitation. Vibration may be caused by the same factors, plus the higher
bearing loads and greater shaft deflection that occur at low flows.
Installation, Operation, Maintenance, and Repair
249
8. Power Savings, Motor Load
As discussed in Chapter 2, Section VII, the BHP curve for higher specific
speed pumps may be flat, or may even slope upward at reduced capacities.
Operating a pump with high specific speed to the left of BEP may actually
cause the motor load to be higher than at BEP.
C. Preferred Operating Range
Refer to the discussion in Chapter 2 on the Preferred Operating Region for
pumps. In that section, it was mentioned that one rule of thumb is to keep
pumps operating within a range of 70%–120% of the BEP flow for continuous
service to maintain stable flow. A good part of the lower end of that range
is based on concerns about recirculation, discussed in Section III.B.5 above.
With higher suction specific speeds, this range becomes even tighter. For
example, if the suction specific speed is above 12,000 or 13,000, the recommended range for stable operation would be tightened to 80%–110% of the
BEP flow. And for very high suction-specific speeds (e.g., using an inducer,
described in Chapter 4, Section II.C, which may result in a suction specific
speed in the range of 18,000–20,000), the recommended range around the
BEP for stable operation may be as tight as ±5% of the BEP flow.
It is important to understand that there is not complete agreement within
the pump industry on these parameters. A related point is that the recommended range around BEP for stable operation depends on whether the
operation is continuous or intermittent, as the recommended range might
be relaxed for intermittent service. Even these terms, continuous or intermittent, are not precisely defined within the pump industry. Some say that operating a pump more than 25% of the time would be considered continuous
duty, while anything less than 25% would be considered intermittent duty.
However, if a pump were to operate around the clock (24/7) for three months
in a year, and then be idle for the remaining nine months, this would be
25% operation, but would it be considered intermittent? The author believes
this should certainly be considered continuous duty. Similarly, operating 6 h
straight during a 24-h period would only be 25% operation, but again, the
author considers this to be continuous duty. In the author’s opinion, anything more than 2 h of continuous operation out of 24 h should be considered
continuous duty.
Note that some pump manufacturers claim that modern design techniques
using computation fluid dynamics (CFD) analysis allows the stable range of
operation around the BEP to be expanded in newer designs. So, the claim
is that lower NPSH (higher suction specific speed) is achievable without so
narrowly restricting operation around BEP. Although this is true in theory,
it may require a sacrifice of efficiency at the BEP to achieve it, so many new
designs do not achieve this wider stable flow range. Also, significant numbers of new hydraulic designs are not introduced every year, so there will
250
Pump Characteristics and Applications
continue to be older hydraulic designs in the field for many years to come.
Consequently, the issues raised in this section will continue to be important.
D. Ten Ways to Prevent Low Flow Damage in Pumps
Centrifugal pumps have a minimum operating flow rate, below which the
pump should not be run for long periods without sacrificing reliability.
Naturally, it is best, from the standpoint of long-term reliability and operating efficiency, to operate pumps close to their BEP, but there may be periods
where reduced flow demand or system changes cause the pump to run at
reduced flow rates. Minimum flow is usually expressed as a percentage of
the flow at the best efficiency point (BEP) of the pump, for a given impeller diameter. Extended operation below recommended minimum flow can
lead to excessive vibration, impeller damage, and premature bearing and
seal failures. With most sealless (magnetic drive and canned motor) pumps,
the allowable runtime below the minimum flow rate may be only a matter of
minutes before significant damage to the pump occurs.
The recommended minimum flow rate varies considerably from one size
and type of pump to another, ranging from 10% to 90% of the BEP flow. The
major characteristics of the pump that influence the determination of minimum flow were discussed in the previous section, and include the energy
level (horsepower per stage) of the pump, the specific hydraulic design of the
impeller inlet, the mechanical design of the pump shaft and bearing system,
and the cost and criticality of the pump. For sealless pumps, the amount of
heat generated in the canned motor or across the magnets is also a consideration, as is the specific heat of the pumped fluid.
There is no accepted industry standard for minimum flow that applies to
all pump types, and even different manufacturers of the same pump type
may have a range of acceptable minimum flows for a given application. That
being said, however, the pump manufacturer is still the best place to start for
the recommended minimum flow for a particular pump installation. Some
manufacturers show this information on the pump performance curve.
What to do, if anything, to protect the pump from the consequences of low
flow damage is an economic decision made by the user. This analysis considers the cost of the pump, minimum flow protection system, and downtime/
lost production, in addition to energy and maintenance costs. Other potential factors may include health, safety, and environmental risks.
A significant majority (estimated at over 80%) of centrifugal pumps have
no minimum flow protection whatsoever. The vast majority of the pumps
installed annually are fairly low horsepower pumps used for transfer or
cooling, which are not normally expected to operate over a wide range of
flow. These pumps are unlikely to operate below the minimum flow point,
except for the inadvertent closure of the main discharge valve or other inadvertent blockage of the system. Furthermore, many of these relatively lowcost pumps are not deemed worth the capital expenditure for minimum flow
Installation, Operation, Maintenance, and Repair
251
protection. This is especially true for noncritical pumps used in residential,
commercial, and light-duty industrial services.
For the 20% or so of pump applications that do require minimum flow protection, there are a number of choices available to the user or system designer. The
determination of which choice to use considers accuracy, reliability, cost, and
criticality, and is very specific to the application. One other factor in the selection process is whether the pump needs to be protected for minimum flow in a
modulating fashion (i.e., keeping the pump operating but with a certain amount
of flow bypassed) or whether it is sufficient to simply alarm or trip off the pump
in the event that the flow rate drops below the recommended minimum flow.
Finally, additional protection obtained from the same instrument should be
considered (e.g., a power monitor can protect against both high and low flow
damage to a pump, while a relief valve will only protect against low flow).
For the 20% or so of pump applications that do require minimum flow protection, the following are ten different methods that can be considered for protection of the pump. All of these are used in industry, and some systems use a
combination of these methods for protection against low flow excursions.
1. Continuous Bypass
This may be the lowest capital cost method of protecting a pump, whereby
a bypass line with an orifice allows a fixed amount of flow to be pumped
continuously back to the suction source. This always ensures that the pump
delivers its recommended minimum flow, even if the main line is shut off
completely. The biggest negative aspect of this system is that the pump must
be oversized in the first place to allow for the continuously bypassed flow.
Second, and sometimes more importantly, is the fact that energy is wasted
due to the extra horsepower required to accommodate the bypassed flow.
There may also be a potential for product damage when being forced through
an orifice. Still, this alternative is chosen by many industrial users for pumps
in the range of 50 HP and below.
2. Multicomponent Control Valve System
This type of system relies on a continuous flow measurement in the system.
When the flow drops below the recommended minimum flow, a signal is
sent to a valve in the bypass line that either opens it completely or modulates
the valve so that it gradually opens. This valve may be a solenoid type if
it is strictly on/off, which is generally the least costly method, or may be a
pneumatically or electromechanically actuated control valve. This method
of bypass eliminates the energy waste of continuous bypass, but relies on
considerably more complexity than a continuous bypass system. The system
includes multiple components, each of which could fail. It requires a power
supply and, if pneumatically actuated, an air supply. Maintenance costs are
typically higher than other alternatives. As such, it is one of the more costly
252
Pump Characteristics and Applications
methods of minimum flow protection. However, it is deemed by many users
to be the best approach, especially if the system already includes a reliable
method of flow measurement.
The operating cost of a pump system with a bypass control valve low-flow
protection system can be much less than that of a system with a continuous
bypass system. In the system shown in Figure 8.7, a 50-HP process pump
(Figure 8.8) must deliver 500 gpm of water for 6000 h annually, and must
operate at the specified minimum flow of 168 gpm for 100 h per year. Power
costs are assumed at 0.10 USD/kWh. Total annual pumping costs are calculated with two minimum flow protection systems: (1) a continuous bypass
and (2) a bypass control valve, which only opens when required. An analysis
of annual pumping costs using these two methods shows that the continuous bypass method costs $3920 more per year to operate. For details on how
this calculation was made, visit the following Web address: http://www.engsoftware.com/kb/item.aspx?article=1310.
3. Variable Frequency Drive
Variable frequency drives (VFDs), described in detail in Chapter 6, Section IV,
change the frequency of the electric motor on the pump to slow the pump down
when the demand for lower flow is called for by the process. For most systems,
this keeps the pump operating near its BEP at all times, and prevents the moving of the pump to a lower percentage of BEP that causes the damage to pumps.
VFDs are being used more and more in process applications and have eliminated the need for other minimum flow protection when they are being used.
Product tank
Bypass flow
control valve
Minimum flow
recirculation line
Supply tank
Flow control valve
set to 500 gpm
Supply pump
FIGURE 8.7
Simple piping system with a minimum flow recirculation line. (Courtesy of Engineered
Software, www.eng-software.com, Lacey, WA.)
253
Installation, Operation, Maintenance, and Repair
200 13 in
55
190
60
65
180 12.375 in
170
Total head (ft)
160
70
72
73
12 in
Impeller trim
needed for
668 gpm
73
73.1
72
150
140
Impeller trim
needed for
500 gpm
130
120
70
110
100
90
9 in
55
80
60
65
70
60
50
50
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900
Flow rate (U.S. gpm)
FIGURE 8.8
The design point on this pump is selected at 668 gpm to deliver 500 gpm through the system and
ensure a 168 gpm continuous bypass. (Courtesy of Engineered Software, www.eng-software.
com, Lacey, WA.)
(Note that with canned motor and magnetic drive pumps, there will still be a
minimum flow required to carry away the heat caused by the motor or magnetic flux and to lubricate the bearings.) VFDs are relatively expensive, although
their cost has reduced dramatically in recent years with rapid improvements in
technology. Other benefits of VFDs include lighter loading of pump seals and
bearings and the ability to “soft start” equipment at slower speeds, thus reducing the strain and high current caused by across-the-line starts.
4. Automatic Recirculation Control Valve
An automatic recirculation control (ARC) valve (Figure 8.9) combines the
features of a check valve, bypass valve, and a flow sensing element, as well
as providing pressure breakdown, and has a number of advantages compared with other approaches. Compared with the multicomponent flow control valve system, it has fewer components, requires lower installation and
operating costs, has less environmental effect (no dynamic seals), and does
not require air or electricity. Compared with systems that just shut down
the pump, it keeps the pump and the system operating (does not shut down
the process). Disadvantages include its relatively high cost, and the fact that
these valves are not normally available in alloys higher than stainless steel,
254
Pump Characteristics and Applications
FIGURE 8.9
Automatic recirculation control valve. (Courtesy of Pentair Valves & Controls, Stafford, TX.)
thus eliminating many chemical services. Also, ARC valves are generally
unsuitable for fluids containing solids.
5. Relief Valve
This system simply relies on a pressure relief valve in the pump discharge
piping being set to relieve back to suction when the pressure put out by the
pump reaches a certain set point pressure. The characteristic performance
curve of all centrifugal pumps is such that as the pump delivers a lower
capacity (flow), the pressure (head) that the pump produces gets higher.
Some pumps exhibit a steeper capacity vs. pressure curve than others. A
pressure relief valve is particularly appropriate for pumps with fairly steep
capacity vs. pressure curves. For example, this is the normally chosen minimum flow protection system for regenerative turbine style pumps. It is also
the accepted method for protection of many fire pump systems. (Note: With
liquid returning to the suction, the internal temperature of the pump may
rise, so this may limit the use of the relief valve for an extended period.) Also,
relief valves are subject to maintenance and testing/calibration regimens.
Finally, it is import to know that most relief valves are not designed to operate continuously (i.e., for a long period), so the pump should not be sized
such that the relief valve is continuously operating.
6. Pressure Sensor
This device relies on the fact that as the flow decreases with a centrifugal pump, the amount of pressure produced by the pump increases. This
Installation, Operation, Maintenance, and Repair
255
high-pressure signal is then used to either open a bypass valve at a highpressure (low-flow) indication or to simply trip the pump. For pumps with
relatively steep head vs. capacity curves, this method can be economical and
reliable. For pumps with flatter head–capacity curves in the low flow range,
it is considered to be less reliable than other approaches.
7. Ammeter
The amp draw of the electric motor varies across the range of flow produced
by a pump. For many pumps, the amp draw of the pump is lower at lower flow
rates and increases with increasing flow. Thus, it is possible with many pump
types to monitor amp draw and to alarm or trip the pump when the amps
drop below a certain set point level. Although this is a relatively inexpensive way to protect the pump against low flow damage, it has some potential
drawbacks. It may be subject to unacceptable inaccuracy due to current fluctuations in the system and the fact that the amp draw curve can be fairly flat
at lower flow rates. In general, the lower the nominal speed of the driver, the
less practical amp monitoring becomes, due to the flatter curve and resulting
smaller amp range. The device may also need to be disabled during startup of
the pump due to the high current draw that occurs then. On the plus side, an
ammeter can also be used to protect a pump from damage due to excess flow.
8. Power Monitor
Power monitors measure motor horsepower. Because most pump curves have a
horsepower curve that rises with increasing flow, it is possible to set the motor
to shut off if the power drops below a minimum setpoint, so this is a reliable
protection against low flow problems. Power monitors are typically more reliable that ammeters because they are not subject to fluctuating results with
variations in line current. For pumps with relatively flat head–capacity curves,
where pressure measurements are not reliable, the power monitor may be
the best choice for low flow protection. Power monitors can be programmed
to protect against excessive flow (high power) as well as minimum flow (low
power). They can also be programmed to ignore momentary power spikes,
where an ammeter might trip the motor. They are adjustable to allow altering
setpoints should the process requirements change. They are not appropriate for many mixed-flow pumps, which may have a nearly flat horsepower
curve as a function of pump flow. If the power monitor measures motor input
power rather than motor output power, it may not be as accurate, because the
efficiencies of small motors at low power can be quite low.
9. Vibration Sensor
Some pump systems have vibration monitors to alarm or trip the pump if the
pump begins to vibrate excessively. One of the things that occurs at lower
256
Pump Characteristics and Applications
flow rates is that the pump may indeed vibrate significantly higher than
normal. Note that high vibration levels may also be an indication of other
problems with the pump, such as misalignment, imbalance of the impeller,
or cavitation. This device, while relatively expensive, is part of the low-flow
protection system on many critical process pumps. If vibration is associated
with pump wear or other factors, such as bearing degradation, it is also possible to project the time of failure and plan preventive maintenance.
10. Temperature Sensor
At very low flow rates, the temperature of the pumped liquid increases due
to friction, the work of compression, and the recirculation of the liquid within
the pump. Thus, if the pump discharge is shut off by a closed main valve, the
temperature of the liquid inside the pump will begin to rise. The rate of temperature rise depends on several factors, including the pump efficiency, the
specific heat of the liquid, pump head, and the volume of liquid in the pump.
One method of protecting the pump against this occurrence is by monitoring the temperature in the pump casing (or containment shell in the case of
a magnetic drive pump), and tripping off the pump when the temperature
rises above a certain setpoint value. This may be relatively inexpensive but
not necessarily too reliable because, by the time it shuts off the pump, damage may have already occurred in the pump.
IV. Maintenance
This section divides maintenance activities into two categories: (1) regular
maintenance is performed according to a fixed schedule, and is done to keep the
pump running optimally, and (2) preventive maintenance has as its specific goal
the prevention of an unplanned emergency shutdown of the pump, which
often results in a much more expensive pump repair when repair becomes
necessary and also often involves costly disruptions in production activities.
A. Regular Maintenance
1. Lubrication
Grease-lubricated pumps should be regreased about every 2000 h (or about
every 3 months for continuous-duty pumps). If the grease cavity is not
vented properly, it may be possible to apply too much grease, causing the
pump to run hot.
Oil-lubricated pump bearing housings typically have the oil level in the
bearing housing set by the manufacturer. Either too high or too low an oil
level can cause the bearings to not be ideally lubricated, reducing their life.
Installation, Operation, Maintenance, and Repair
257
A common system for maintaining a constant oil level in the bearing housing is shown in Chapter 4, Figure 4.12. A leveling bar below the sight glass is
set by the manufacturer at the prescribed oil level. If the level drops below the
setpoint (due to leakage past the oil seals), it creates an air path that allows oil
to flow from the sight glass into the bearing housing until the prescribed oil
level is reached again. Thus, the sight glass serves as an inventory of spare
oil to make certain the oil level stays constant and as a visual indicator to
alert the operator that the oil seals are leaking and need replacement or that
additional oil needs to be added.
Oil mist lubrication of rotating equipment through a centralized plant
oil mist system is becoming more popular. This type of lubrication system,
although requiring capital investment to implement, can result in the benefits of longer bearing life and reduced oil consumption.
Note that some coupling types (e.g., gear and grid spring types) require
periodic lubrication.
2. Packing
Cutting and installing packing is an art, and proper installation techniques can
do a lot to extend the service life of the packing. When making packing rings
from a roll, a dummy shaft of the same size as the pump shaft or sleeve (if the
pump has a sleeve) can be wrapped with the packing to ensure that the rings
are cut to the right length. Cutting the rings on a diagonal rather than straight
across helps minimize leakage through the split (the point where the two ends
of a ring come together when they are wrapped around the shaft or sleeve).
When the rings are installed, each ring should be compressed as it is
installed, with a split bushing that can slide over the shaft or with tampers. If
this is not done, the lower rings of packing are not compressed when the gland
is tightened down, and only the final outer rings of packing do any sealing.
Another good idea is to stagger the splits of the individual rings 90° apart
from each other as the packing rings are installed, to prevent the possibility
of liquid leaking out through the ring splits.
Before installing new packing, it is a good idea to examine the old packing
and the stuffing box, especially if the packing has prematurely failed. This can
sometimes provide a clue as to why the packing failed and can prevent a recurrence of the problem. For example, if the outside surfaces of the packing rings
(next to the stuffing box bore) are worn instead of the inside surfaces (next to the
shaft or sleeve), the packing was likely not installed properly, or else the packing
size used was too small for the stuffing box bore. If the outer rings of packing
(the ones nearest the gland) are worn but the inner ones are not worn, chances
are the inner rings were not compressed by the technician when the packing
was installed, so the outer rings of packing were the only ones doing any work.
Another thing to consider at the time that packing is being replaced is the
possibility of hardening the shaft or sleeve that runs under the packing with
a metallic, ceramic, or other type of hard coating. This minor modification of
258
Pump Characteristics and Applications
the pump can substantially increase the operating runtime before the next
repacking. Excessive leakage around the shaft is often the first thing to go
wrong with a pump, especially if the sleeve is not hardened or if the liquid
contains abrasives.
3. Seals
When doing maintenance on a mechanical seal, care should be taken in handling the seal’s dynamic running faces. Very small impurities, even oil residue on the technician’s hands, can reduce the seal’s performance or life.
Lubricant used to mount O-rings or other elastomeric parts of seals should
be compatible with the elastomeric material. Hydrocarbon-based oils should
not be used for lubricant because these oils can cause some elastomeric materials to swell. A vegetable-based oil such as corn oil or a synthetic lubricant
is generally better for this function because these will not cause swelling of
most seal elastomers.
Set screws, O-rings, and, if possible, springs should be replaced if a seal is
being overhauled.
B. Preventive Maintenance
A preventive maintenance program relies on the following maintenance
operations being periodically performed on pumps to lengthen the runtime
between repairs, limit the damage done to pumps, and prevent unscheduled
outages of the equipment. Each preventive maintenance program is unique
and depends on the size, cost, location, and importance of the pumps; the
size and experience level of the pump owner’s maintenance organization;
and the severity of the pump services.
1. Regular Lubrication
As discussed in Section IV.A.1, regular lubrication can increase the life of the
pump and driver bearings, and the coupling if it requires lubrication.
2. Rechecking Alignment
A quick periodic alignment check when the pump is shut down ensures that
the equipment maintains good alignment and does not move out of alignment due to vibration or structural shifts.
3. Rebalance Rotating Element
This should be done during a pump overhaul if the impellers or coupling are
machined or otherwise modified or if they have suffered any wear, erosion,
or cavitation damage.
Installation, Operation, Maintenance, and Repair
259
4. Monitoring Benchmarks
The three most important benchmarks to monitor as part of a preventive maintenance program are the (1) pump hydraulic performance, (2) temperature,
and (3) vibration. These three benchmarks are discussed in more detail below.
C. Benchmarks
1. Hydraulic Performance
Chapter 3, Section VI, discusses procedures and instrumentation for measuring pump capacity, total head, and power in an installed system. These data
should be recorded when the pump is first put into service to the best extent
possible with the testing instrumentation that is available at the installation.
If no flow measuring instrumentation is available at the site, pump total head
(or discharge pressure if the liquid density will not change) can be measured
with the discharge valve in its normally open position and fully closed.
(Check with the manufacturer, however, before testing with a fully closed
discharge valve, as some pumps cannot be operated, even briefly, with the
discharge valve fully closed.)
When the hydraulic performance parameters are periodically rechecked,
this provides an indication of the deterioration of pump performance caused
by the wear in the leakage joint, other corrosive and/or erosive wear on the
pump impeller or casing, and cavitation.
2. Temperature
A pump begins to run hotter if its bearing is about to fail, if it has inadequate
lubrication or cooling, or if the pump is misaligned. These are excellent reasons
to monitor the temperature of the pump at the bearing housing and to investigate any increase in bearing temperature. Many engineers and maintenance
technicians make a habit of feeling the bearing housing with their hands to
see if the bearings are running hot. Unfortunately, the human hand does not
make a very good thermometer, as anything over about 130°F feels “too hot” to
the hand, while pump bearings can run much hotter than this. The upper temperature limit is when the oil or grease lubricant begins to carburize, which is
in the range of 180°F to 200°F for most lubricants. A much better indication of
the pump bearing temperature can be obtained using a thermocouple, a thermometer in a thermowell, or other temperature-recording instrument.
If the pump bearings appear to be running hot, consideration should be
given to simple reasons why the pump might be running hot before pushing
the panic button. The oil level in the bearing housing should be checked to
make certain that the bearings have adequate lubrication. Perhaps the process
liquid is hotter than usual, which would conduct heat along the shaft and cause
the pump bearings to run hotter. Maybe the pump has lost cooling water to the
bearing housing or seal, or perhaps the cooling water temperature is higher
260
Pump Characteristics and Applications
than normal. Consider the ambient temperature, because a pump tends to
run hotter when the ambient temperature is hotter. If none of these considerations apply, a vibration check (discussed below) should be made prior to
shutting down the pump. After the pump is shut down, the cause of the high
temperature can be further investigated. The direction of rotation and alignment should be checked first, and following that, other possible causes can be
considered using the troubleshooting techniques discussed in Section V below.
Note that a new or newly overhauled pump should be run for an hour or so
(at operating temperature) before establishing the temperature benchmark.
It can actually run a little hotter when the pump is first started, until the oil
seals run in a bit.
3. Vibration
Figure 8.10 shows the recommended maximum acceptable field vibration of
horizontal end suction frame-mounted pumps handling clear liquids. The
figure shows maximum vibration velocity in inches per second RMS (and
above that, in millimeters per second RMS), unfiltered, as a function of input
power. This chart shows, for example, that a pump with an input power of
10 HP (7.5 kW) has a maximum allowable field vibration of about 0.14 in/s
(3.5 mm/s), whereas one with an input power of 100 HP (75 kW) has a maximum vibration allowance of about 0.21 in/s (5.3 mm/s). Vibration velocity is
measured on the bearing housing of horizontal pumps and at the top motor
bearing of vertical pumps. Measurements should be taken in all three planes
(H, V, and A in Figure 8.10), with Figure 8.10 showing the maximum allowable vibration measured in any plane.
Some pump users are reluctant to check vibration readings on pumps.
Their argument is that because they do not have the sophisticated multispectrum vibration analysis equipment that measures frequency as well as displacement or velocity of vibration, they are not able to determine the cause
of the vibration. It is true that to perform a thorough analysis to determine
if vibration is being caused by imbalance, misalignment, bad bearing, etc.,
one must have instrumentation to determine the frequency as well as the
amplitude or velocity of vibration. However, a simple measurement of vibration displacement or velocity (comparing it to the benchmark reading taken
when the pump was first put into service) is enough to call attention to the
fact that a problem is developing. The user can then bring in more sophisticated analysis equipment or do other checks to help determine the root cause
of the problem, hopefully before long term damage is done to the pump.
A simple handheld probe to measure vibration displacement or velocity is
adequate to serve as a preventive maintenance tool for setting a benchmark
level and comparing periodic readings against this value.
If more sophisticated vibration equipment is available (along with the
proper training on its use), vibration readings can be used to do a much
more in-depth troubleshooting analysis of a pump. Table 8.1 is a vibration
261
Installation, Operation, Maintenance, and Repair
Vibration (mm/s RMS, unfiltered)
7.11
6.10
5.08
4.06
3.04
2.03
1.01
0
0.75
7.5
75
750
Input power at test conditions (kW)
Vibration (inches/s RMS, unfiltered)
0.28
0.24
0.20
0.16
0.12
0.08
0.04
0
1
10
100
Input power at test conditions (BHP)
1000
V
A
H
FIGURE 8.10
Acceptable field vibration limits for centrifugal pumps handling clean liquids. (Courtesy of the
Hydraulic Institute, Parsippany, NJ; www.pumps.org.)
1×
Less than 1×
0.6× to 0.93×
Vibration may increase dramatically shortly after this appears
Exactly 1/2×,
1/3×, 1/4×, etc.
0.41× to 0.49×
Smaller peaks at (1 – f)×, and at ± (1 – f)× “sidebands” of the first several
multiples of running speed (where f = 0.6 to 0.93)
This is often accompanied by rumbling noise and “beating” felt in the
foundation
It occurs at part-load capacities, but disappears at very low capacity and
independently depends on speed and flow
Increased broadband vibration and noise level below running speed as
NPSH decreases, especially at high flows
Often accompanied by cracking noise
Stronger on shaft than on housing. Hydraulic performance and/or suction
pressure normal. Axial vibrations within normal limits and vibrations
above runout increase with roughly speed squared
Vibration highest on driver/driven IB bearing housings
Near shaft natural frequency and orbit “pulses,” forming an inside loop
(a smaller orbit inside of a larger orbit)
Vibration onset is sudden at a speed roughly twice the excited natural
frequency, and “locks” onto the natural frequency despite speed increase
Vibration response is broad in frequency
Other Symptoms
0.05× to 0.35×
Multiples of
Running Speed
Vibration Troubleshooting Summary Chart
TABLE 8.1
Pump coupling imbalance
Imbalance in rotating assembly
Classical cavitation (i.e., cavitation without
recirculation)
Diffuser or return channel stall (flow vortices
around the front of the diffuser)
Light rub, combined with low shaft, or bearing
support natural frequency
Stable fluid whirl if whirl frequency is below
critical speed, or rotordynamic instability due to
fluid whirl in close clearances if whirl frequency
becomes equal to shaft critical speed (e.g., bearing
“oil whip”)
Internal flow recirculation probably at suction,
particularly if accompanied by the crackling noise
of cavitation at flow rates below BEP
Probable Causes
262
Pump Characteristics and Applications
2×
Shaft vibrations much stronger than housing vibrations, and approach
or exceed bearing clearance. Decrease in shaft first bending natural
frequency. Other multiples of running speed may be stronger than
usual, but especially 2×
Shaft vibrations stronger than housing, and torque pulses
In motor-driven pumps, with speed equal to electric line frequency,
and highest vibration at motor IB housing, with vibration exactly at
2× line current frequency of 50 or 60 Hz
Axial vibrations are over 1/2 of horizontal vibrations, 1× vibrations are also
high, and bearing oil temperature is high
Vibration highest on driver IB or OB bearing housing
Vibration high on pump IB or OB housing, low on driver
Natural frequency near 1×, as determined by bump test
Axial vibrations are over 1/2 of H or V vibrations, or vibrations increase
much slower than the square of the speed. Also, bearing oil temperature is
high, and/or temperature readings show one side of the bearing housing
is hotter than the other by more than 5°F
Discharge pressure pulsations are strong at 1× but not at impeller vane pass
in a single volute pump
Discharge pressure pulsations are strong, particularly at vane pass,
especially at flows far above or below the design point
Axial vibrations are low
Both shaft and housing vibrations are strong, and discharge pressure
pulsations are strong, but impeller vane pass vibrations are low in a
twin volute pump
Same as above, but with high vane pass vibrations and discharge
pressure pulsations
(continued)
Pump/driver misalignment at the coupling
Torsional excitation
Electrical problem with motor
Twin volute vanes designed too close to impeller
OD, or clogged or damaged volute, or impeller/
volute eccentricity
Looseness in bearing support or cracked shaft
Clogged or damaged impeller passage
Volute tongue designed too close to impeller OD, or
excessive impeller/volute eccentricity
Clogged or damaged impeller passage
Driver rotor imbalance
Pump rotor imbalance
Resonance
Pump/driver misalignment at the coupling
Installation, Operation, Maintenance, and Repair
263
Other Symptoms
Discharge pressure pulsations reasonably low and both shaft and housing
vibrations high
Discharge pulsations relatively low, but housing vibrations much higher
than shaft vibrations
Discharge pressure pulsations high at vane pass frequency but suction
pressure pulsations reasonably low
Suction pressure pulsations high
Pump rotor or casing natural frequency close to vane pass
Orbit shows sharp angles or shows evidence of “ringing,” and/or spectrum
shows evidence of exactly 1/2 or 1/3× response; grinding noises and
speed changes may be evident
Orbit is “fuzzy” but does not pulse or “ring”; spectrum may also exhibit
1/3 or 1/2 running speed
Orbit pulses, usually in one direction much more than the other and shaft
vibrates more than housing; harmonics of exactly 1/2× may also be present
Housing and casing vibrate much more than shaft, combined with vibration
response over a broad range of frequencies below running speed
Probable Causes
Shaft support looseness, especially of bearing to
bearing housing or bearing housing to casing
Looseness in pump casing, pedestal, or foundation
Jammed, clogged, or damaged seal
Acoustic resonance in suction pipe
Resonance
Internal rub or poorly lubricated gear coupling
Acoustic resonance in discharge pipe
Volute too close to impeller OD due to design or
excessive rotor eccentricity
Piping mechanical resonance at vane pass
Source: Courtesy of William D. Marscher, Mechanical Solutions, Inc., Parsippany NJ, www.mechsol.com.
Several multiples
of running
speed,
including
1×, 2×, 3×, 4×,
and possibly
higher
Number of
impeller vanes ×
running speed
Multiples of
Running Speed
Vibration Troubleshooting Summary Chart
TABLE 8.1 (Continued)
264
Pump Characteristics and Applications
Installation, Operation, Maintenance, and Repair
265
troubleshooting summary chart that can serve as a guide to vibration root
causes, and how to interpret vibration symptoms. It is not meant to include
all possibilities and is in the order of the frequency value observed, not in
order of likelihood or importance to reliability.
In studying Table 8.1, be mindful of the following definitions:
• Runout: false vibrations picked up by a proximity probe, actually
reflecting shaft scratches, gouges, undulations, etc., or static shaft
bends or misalignment (“mechanical runout”), or eddy current
sensitivity variations along the shaft surface (“electrical runout”).
Runout can be quantified by observing the apparent shaft orbit with
a dial indicator when the shaft is slowly turned.
• Inboard (IB): the coupling side of the pump or driver.
• Outboard (OB): the end of the pump or driver opposite to the
coupling.
• Narrowband: vibration response that is at a single frequency.
• Broadband: vibration response that covers a band of frequencies on a
spectrum plot.
V. Troubleshooting
A typical on-the-job problem faced by readers might be an operating pump
making a lot of noise, or with bearings running too hot. These symptoms
could be due to any of a number of causes. They might be caused by
vibration due to imbalance, vibration due to misalignment, cavitation, air
entrainment, or a number of other possible causes. Troubleshooting such
problems is like detective work. The problem is solved by making a list of
suspects and trying to eliminate the suspects one at a time, starting with
the ones that are most easily eliminated. The ones most easily eliminated
are usually those that can be checked out without shutting down the
pump. The next ones to be eliminated are those that can be analyzed by
shutting down the pump, but without having to disassemble the pump.
Finally, the last group of suspects are those that can only be checked by
disassembling the pump.
Table 8.2 is a troubleshooting chart that lists eight of the most commonly
observed symptoms of centrifugal pump problems, along with the possible
causes for these symptoms.
266
Pump Characteristics and Applications
TABLE 8.2
Pump Troubleshooting Chart
Symptom
Possible Causes
Insufficient flow
Pressure too low
High amp reading
Packing leaks too much
Seal/packing fails early
High vibration/noise
Bearings short lived
Bearings run too hot
1-2-3-4-6-7-10-11-12-13-20
1-2-3-4-6-9-10-11-12-14-20
5-6-7-8-9-10-13-15-17-18-22-23-26
15-17-21-22-23-24-25-27-28
1-7-11-15-17-19-21-22-23-24-25-26-27-28
1-2-3-7-8-11-13-15-16-17-18-19-20-24-25
1-6-7-9-11-15-17-24-25-28-29-30
1-7-11-15-18-19-24-25-29-30-31-32
Causes
1. Air entrainment
2. Suction obstructed
3. Poor sump design
4. Speed too low
5. Speed too high
6. Incorrect rotation
7. Pump flow too low
8. Pump flow too high
9. Change in density
10. Change in viscosity
11. Cavitation
12. Leakage joint excessive
13. Foreign matter in impeller
14. Loose impeller
15. Misalignment
16. Foundation not rigid
17. Shaft bent
18. Impeller or wear ring rubbing
19. Worn bearings
20. Damaged impeller
21. Shaft scored at packing/seal
22. Packing installed improperly
23. Incorrect type of packing
24. Excessive shaft runout
25. Impeller or coupling unbalanced
26. Gland too tight
27. Stuff box bushing clearance high
28. Dirt or grit in pumped liquid
29. Bearing cooling water failure
30. Inadequate bearing lubrication
31. High ambient temperature
32. High process liquid temperature
VI. Repair
A. General
A thorough discussion of pump repair would fill another entire book and is
therefore beyond the scope of this section. Some very light-duty pumps such
as residential pumps have very little repair capability and must simply be
replaced when they wear out. Commercial and industrial pumps are generally repairable to some degree or another, with the degree of complexity and
therefore the amount of repair that can be economically done varying considerably. Some sophisticated pump users maintain elaborate repair shops to
do their own pump repairs, whereas other users send out their equipment to
repair centers that are especially suited to the repair of rotating equipment.
If it is available, the manufacturer’s instruction manual should be consulted prior to beginning any repair operations.
Whether the pump is to be repaired in the owner’s own shop or in an outside
repair center, it is good to remember that not all pump repairs are equal. A
Installation, Operation, Maintenance, and Repair
267
very minimal repair job might not include any checking of the fits and running clearances in the pump. It might ignore any analysis of what went wrong
to cause the pump to require repair. It might merely consist of changing a few
very standard components (such as bearings and seals) with factory parts.
At the other extreme, a much more thorough repair procedure might
include a careful measurement and recording of all fits and clearances (and
a repair of any component that is out of specification in this regard); a check
of shafts for straightness; a rebalancing of all impellers after any welding,
grinding, or machining operations have been completed; a concentricity
check of all bearing housings, stuffing boxes, and similar components; etc.
Obviously, the more thorough repair job takes longer and is more expensive than the “quick-and-dirty” one. Hopefully, however, the more thorough
repair job results in a longer runtime before the pump must be overhauled
again. With the high cost of pulling the pump and placing it back in service,
and the high cost resulting from plant operations being curtailed while the
pump is being repaired, many users find that the extra money spent to do a
more thorough pump repair is money well spent.
Just as there are different levels of completeness or accuracy of pump repair
that can affect the longevity of the pump between repair cycles, users should
also be aware that there may be different levels of quality in the replacement
parts being used. Parts are usually available from the original maker of the
pump, but they may also be available from non-OEM parts makers, or some
parts may be able to be made in the owner’s shop. Not all machine shops have
the same levels of accuracy in their machine tools, have the same skill level in
machinists, or have available the correct material and latest drawing for the
part. This is particularly important in the case of cast parts of irregular shape,
such as impellers and volutes, where deviations in dimensions can affect the
pump’s hydraulic performance as well as its operability and longevity.
As a final general suggestion on pump repair, the pump owner should
maintain as thorough a repair history on each pump as possible, complete
with the documentation from each repair (see Section VI.B.1 to follow) and
repair parts inventory. Accurate records help when making decisions on
when repair operations should be done, speed up the repair, reduce the cost
of parts and repairs, and may allow a better analysis of the reasons for overly
frequent pump repair so that design modifications can be implemented to
lengthen the operating time between repairs.
B. Repair Tips
Below are a few important practices to follow during the repair of pumps.
1. Document the Disassembly
The disassembly of a pump to be repaired should be well documented.
Components arranged in a particular order on the pump (such as the bowls
268
Pump Characteristics and Applications
on a multistage vertical turbine pump) should be match-marked or punched
with a number to ensure that the pump is reassembled with the components
in the same order. Remove any extraneous marks to avoid confusion.
Key dimensions should be measured with micrometers after the components have been disassembled and cleaned. This should include all fits and
running clearances, so that these can be compared with the manufacturer’s
recommended clearances, or with the user’s own standards, and replaced or
reworked if they fall outside the standard. (See further comments on fits and
clearances in Section VI.B.5 to follow.)
Make sketches or photographs of the components if a drawing is not available, showing clearly where the measurements have been taken and indicating particularly worn or damaged parts.
2. Analyze Disassembled Pump
The disassembled pump should be thoroughly inspected and analyzed, particularly if the user believes that the length of time since the last overhaul is
unusually short. A thorough inspection and analysis of the disassembled
pump can have the major benefit of reducing the likelihood of pump failure
reoccurring so quickly from the same cause.
The inspection of the disassembled pump is an attempt to determine the
root cause of the pump failure, if that is not already known, by the nature
of the damage to the components. The second question to answer through
the analysis of the pump prior to beginning repair operations is whether
the runtime of the pump can be economically extended through a design
modification or a particular repair operation. Examples of design modifications that might be considered to increase the runtime of the pump include
hardfacing or coating of shafts or sleeves at bushing locations or under the
packing; changing the material of key wear components such as wear rings,
sleeves, or bushings; switching from packing to a mechanical seal or changing seal type or materials; installing heavier-duty bearings; and applying
coatings in the casing at areas subject to excessive wear.
3. Bearing Replacement
Bearings are usually replaced during a pump overhaul. Radial and ball bearings commonly used in pumps often fail due to the presence of dirt, moisture, or other debris. Consequently, every effort should be made to keep the
bearing and its environment as clean as possible during the repair operation.
Keep new bearings enclosed in their protective wrapping paper and box
until they are ready for installation. Keep the bearings covered after they
have been unwrapped until they are installed. The mechanic who changes
the bearings should have reasonably clean hands and a clean workbench.
Carefully inspect the bearing housings for dirt, machine cuttings, etc., and
clean the bearing housings with an appropriate solvent or cleaning agent.
Installation, Operation, Maintenance, and Repair
269
Thoroughly inspect the shafts on which the bearings are to be mounted,
removing any burs or nicks with emery cloth, and check to ensure that the
shoulders on which the bearings will rest are square. Finally, the new bearings themselves should be thoroughly inspected for foreign materials or
moisture, flat spots or rust on balls or rollers, or any other visible problems.
The tolerances between the inner race of bearings and the shaft and
between the outer race and the bearing housing are quite tight. Therefore, be
very sure that bearings are oriented properly before mounting them on the
shaft and in the housing. Many bearings have the shields arranged in such a
way that there is a correct way and an incorrect way for them to be oriented.
A bearing that is incorrectly mounted and must be pulled usually is out of
tolerance and has to be tossed.
If the repair facility permits, it is better to heat bearings with oil or an
electric heater prior to mounting them on the shaft. The heating expands the
inside diameter of the bearings, allowing them to slip over the shaft easily.
If the bearings are not heated, they should be pressed on the shaft very carefully, using a bushing so that the bearings are pressed evenly from all sides
and are not allowed to cock.
Finally, make certain when bearings are installed that they are tightly
pushed up against the shoulder on the shaft, that they are properly locked
in place with the snap rings or lock nuts if there are any, and that they are
lubricated properly (or tagged for lubrication as discussed in Section VI.B.9
to follow).
4. Wear Ring Replacement
Recommended new or rebuilt wear ring clearances are discussed in Chapter
4, Section II.A. Good maintenance practice is to replace wear rings or other­
wise renew the clearance when the clearance reaches twice the original
amount. Alternatively, the decision to renew clearances can be based on a
determination through testing of the energy being wasted by having excessive wear ring clearances.
Before removing old wear rings, check and record the worn wear ring
clearance and compare with the specification for the original ring clearance.
Especially for large pumps or pumps with expensive wear ring materials, it
may be possible to replace only one part of the wearing ring (casing ring or
impeller ring only), truing up the ring that is going to be used again, and then
machining a new mating ring to have the proper ring clearance. This technique
requires wear ring components to be purchased or made with undersized
inside diameters on casing rings and oversized outside diameters on impeller
rings and requires careful measurement and record keeping. However, it may
save the user money on wear ring replacement in the long run.
Before removing old wear rings, remove any set screws that are holding the rings in position. Most old wear rings, which are usually mounted
with a press fit in the casing and an interference fit on the impeller, can be
270
Pump Characteristics and Applications
simply pulled off. If that does not work, they may need to be machined off.
Alternatively, casing rings in certain alloys such as stainless steel can be
removed by running a weld bead around the inside of the ring. This causes
the ring to shrink, allowing it to be removed by hand.
After new wear rings have been mounted and pinned in place, the concentricity of the impeller wear ring should be checked by mounting the impeller
on the shaft or a mandrel, mounting the shaft in a lathe or on rollers, and
checking the runout on the outside diameter of the impeller wear ring with a
dial indicator. By performing this operation in a lathe, the ring can be given
a final skin cut if it is not running concentric. The ring is especially likely to
have a problem with concentricity if it is mounted with axially oriented set
screws, with the ring being likely to dimple at the set screw locations.
5. Guidelines for Fits and Clearances
If the pump manufacturer or owner does not have established criteria that
differ from these, the guidelines shown in Table 8.3 may be used for fits and
clearances in centrifugal pump repair. These are conservative guidelines for
industrial pumps and can be relaxed on pumps used in lighter-duty services.
6. Always Replace Consumables
All gaskets, O-rings, set screws, springs, lip seals, packing, and lubrication
should be replaced any time a pump is repaired. Mechanical seals should be
replaced or rebuilt. The dynamic seal faces of many mechanical seals can be
re-lapped to restore their flatness and surface finish.
TABLE 8.3
Fits and Clearances for Centrifugal Pump Repair
Component Description
Fit or Clearance
Ball bearing i.d. to shaft
Ball bearing o.d. to housing
0.0001–0.0007 in interference
0.0005–0.001 in clearance; bush housing if
> 0.0015 in oversize
0.001–0.0015 in clearance
Metal-to-metal to 0.0005 in clearance
See Chapter 4, Section II.A
0.002–0.003 in interference
0.002–0.003 in interference
0.015–0.020 in clearance
Metal-to-metal to 0.0005 in clearance
0.002–0.003 in clearance
Flat within three light bands
0.004 in maximum clearance; weld pads
and re-machine if oversize
Sleeve to shaft
Impeller to shaft
Wear ring to wear ring
Impeller ring to hub and case ring to casing
Throat bushing to case
Throat bushing to shaft
Coupling to shaft
Deflector to shaft (set screw mounted)
Seal running faces
Alignment fits for case, covers, bearing
housings, stuffing box
Installation, Operation, Maintenance, and Repair
271
7. Balance Impellers and Couplings
In general, the major rotating components of pumps (impellers and couplings) should be balanced to maintain the pump within acceptable vibration levels. The need for balance, type of balance, and limits of imbalance
vary with the weight of the component and its radius as well as the operating
speed of the pump. Refer to the manufacturer’s recommendations for each
pump. Many couplings do not require balancing, and smaller impellers may
only require a static balance. This is done by mounting the impeller on a shaft
on rollers, and seeing whether one side of the impeller tends to rotate to the
bottom. Material is then removed from this heavier side.
A dynamic balance is performed on larger impellers, and some industrial
users do a dynamic balance on all impellers of repaired pumps. Dynamic
balance is performed using special balancing machines, with the pump
mounted on a shaft or mandrel.
8. Check Runout of Assembled Pump
This check of shaft runout on the fully assembled pump is a final assurance that all components are machined accurately, that the shaft is straight,
and that the pump has been aligned properly. Mount a dial indicator on the
pump bedplate using a clamp attachment or magnetic base and locate the
pin of the indicator on the shaft at the mechanical seal or packing. Turn
the shaft by hand at the coupling and note the total runout in one revolution.
Shaft runout should be as low as possible. For comparison with a common
standard, the API 610 Standard (Chapter 4, Section XIV.C) calls for a maximum shaft runout of 0.002 in. This is a tight standard; so for pumps not built
to API 610, this standard can be relaxed to 0.003 to 0.004 in, but 0.002 in is a
worthy goal.
9. Tag Lubrication Status
If the pump or motor requires lubrication prior to being run (and most do),
the pump should be tagged to indicate the status of the lubrication (i.e., if
the lubrication was done in the shop during the overhaul, and what type
of lubricant was used). If the equipment has no lubrication, an incorrect oil
level, or requires relubricating prior to startup, it is especially important to
indicate this on a tag to prevent damage to the pump or motor.
10. Cover Openings Prior to Shipment
This prevents debris from getting trapped inside the pump during transit
from the repair shop to the field.
9
Case Studies
I. Introduction
This chapter consists of a number of case studies discussing actual pump
problems with which the author has dealt over the years, or which are based
on contributions from students or colleagues of the author. All of the case
studies are based on historically accurate situations, though changes have
been made in the identity of the pump owner, equipment manufacturers,
and individuals involved, as well as some nonessential details of the cases,
to preserve the confidentiality of event and the anonymity of the people and
companies involved.
Every effort has been made to provide a wide range of pump case study
topics, including pump hydraulics and systems (e.g., poorly sized pumps,
NPSH problems), incorrect pump operational procedures, and problems
that occurred as a result of incorrect or insufficient pump maintenance and
repair activities. The wide range of businesses and pump types represented
in these case studies serves to illustrate the similarity of many pumping
problems across a wide spectrum of industries and pump installations. Very
few pump problems are unique to a single industry or application, and there
are lessons to be learned from each of these case studies.
II. Case Studies
A. The Case of the Oversized Pump
1. Background
The Hydrogen DeSulfurizer (HDS) unit at the Reliable Oil Company’s
Omega Refinery was originally constructed in 1975. The HDS unit was
equipped with two Light Ends Pumps that were designed to deliver 60 gpm
at a total head of 180 ft. These were single-stage overhung pumps built
to API 610 standards. (Refer to the discussion on the API 610 standard in
© 2008 Taylor & Francis Group, LLC
273
274
Pump Characteristics and Applications
Chapter 4, Section XIV.C.) The pumps were driven by electric motors at
3560 rpm. One pump was always in operation and the other pump served
as a standby.
When the original pumps were badly damaged by a fire in the refinery
in 1980, they were replaced with pumps that Omega Refinery had in their
warehouse. The replacement pumps are capable of producing a higher head
and flow than the original pumps. The new pumps are designed to deliver
230 gpm at a total head of 325 ft. These pumps are also single-stage overhung
pumps that are driven by electric motors at 3560 rpm. As with the original
pumps, one replacement pump is always in operation and the other pump
serves as a spare.
Since the replacement pumps were installed, these pumps have had a history of frequent mechanical seal failures, rarely lasting more than 9 months
between seal failures. Light Ends Pump A has just experienced yet another
mechanical seal failure, and the reliability engineering group has been given
the assignment to investigate the causes of these failures and make corrections to address the problems.
2. Analysis of the Problem
The operating conditions are obtained from the pump data sheet and confirmed to be correct by the unit Process Engineer. One data point that raises
a flag is the fact that the liquid vapor pressure (165 psia) at the pumping
temperature is very nearly the same as the pump suction pressure (150 psig,
roughly 164.7 psia). A cross section drawing of the pump is examined, and
one problem is identified. The pumps were built with integral seal chambers that are flushed using an API Plan 1. (Refer to the discussion of seal
piping plans in Chapter 5, Section IV.C.6. API Plan 1 is an internal recirculation from the discharge area of the pump into the seal chamber, similar to a Plan 11, but with no exposed piping.) The API Plan 1 directs the
flush, at near discharge pressure (240 psig), to the seal chamber. However,
the seal chamber in this pump is operated at a pressure that is very close
to suction pressure, because of the thrust balance recirculation holes in the
impeller. (See Chapter 4, Figure 4.5, for an illustration of this thrust balancing arrangement.) Although the flow of the flush can remove heat from the
seal faces, there is not enough pressure in the seal chamber of this pump to
provide an adequate margin above the vapor pressure of the product. A 30
to 40-psi margin above the vapor pressure of the product is desired to maintain proper lubrication of the seal faces. Failure to maintain this margin will
result in a premature failure of the mechanical seal.
The original design flow rate for these pumps was listed as 230 gpm, as
noted above. A trend of the flow rate through these pumps is retrieved from
the process computer for comparison against the design flow rate, which
reveals another problem. The actual flow rate through these pumps ranged
from 30 to 90 gal/min. The BEP for the impeller being used on the pumps is
© 2008 Taylor & Francis Group, LLC
Case Studies
275
300 gal/min. The actual flow rate through these pumps equates to 10%–30%
of BEP. To make matters worse, the impellers in these pumps have a suction specific speed of 11,850, which indicates that the stable operating range
for these impellers is narrower than for many other pumps. (See Chapter
8, Section III.B for a discussion of the narrower operating range required
on many pumps with higher suction specific speed impellers.) The actual
operating range for these pumps falls well below the stable operating range
for the impeller.
In summary, two significant problems are identified as the probable cause
of the failures of the mechanical seals on the pumps.
• The margin between the seal chamber pressure and the vapor pressure is not sufficient to provide proper seal face lubrication.
• The pump is operating well below its stable operating range, resulting in increased shaft deflection, bearing loads, and recirculation
vibration.
3. Solutions and Lessons Learned
The reliability engineering team realizes that the main cause of the repeated
pump failures was that the pumps that replaced the original pumps in 1980
were grossly oversized for the intended service. The replacement pumps
were used because they were available in the Omega Refinery warehouse,
and would fit into the piping system with only minor casing and bedplate
modifications. The replacement, in hindsight, was done with very little analysis of actual system service conditions.
To solve the problems with these pumps, the reliability engineering team
considers three options:
• Replace the pumps completely with new pumps that would be properly sized and designed for the application.
• Upgrade the power frames and mechanical seals and hydraulically
re-rate both pumps. A hydraulic re-rate usually consists of replacing a pump impeller with a new impeller that will fit in the original
pump casing, which also may have to be modified. It is commonly
considered in refineries and other similar process services where
new pumps are quite costly and may have very long lead times.
• Recirculate the pumped product to maintain the required minimum
flow through the pump.
The third option above is not available, as there is no minimum flow system installed for these pumps. Installation of a minimum flow system would
require the involvement of other engineering disciplines (Fixed Equipment
and Instrumentation). This work would likely be very low on their priority
© 2008 Taylor & Francis Group, LLC
276
Pump Characteristics and Applications
list and would most likely not get done for some time. Installation of the
minimum flow system would also be very costly.
With the first option above estimated to be by far more costly than the
second option, the reliability engineering group elects to move forward with
the second option, the hydraulic re-rate and upgrade of the seal and power
frame.
An impeller is chosen from the pump manufacturer’s catalog that would
fit into the existing pumps and which yields a flow rate much closer to the
actual system requirement, though it is still larger than necessary. The
expected operating flow range (30–90 gpm) equates to 27%–82% of the new
impeller’s BEP. The suction specific speed for this new impeller is 6600,
which indicates that the impeller has a very wide stable range of operation.
Since the pump is only operated at the lower end of the flow range for short
periods, selection of this impeller is not expected to present any significant
problems, even though it is not optimal in terms of efficiency.
The pump volute is reshaped so that the pump will be able to develop
the desired head–capacity curve. Figure 9.1 shows how the pump volute
is reshaped, typical of casing redesigns that often accompany hydraulic
re-rates. Since the new impeller has a much smaller eye than the one it is
Volute
insert
FIGURE 9.1
The volute insert shown here allows the re-rated pump to develop the desired head–capacity
performance curve.
© 2008 Taylor & Francis Group, LLC
Case Studies
277
replacing, the case wear ring is replaced with a redesigned wear ring. The
new case wear ring provides a smooth transition for the flow from the pump
suction nozzle into the eye of the impeller.
Several modifications are made to the mechanical seal chamber to raise the
pressure in the seal chamber and provide more uniform seal flush. A close
clearance bushing is installed in the bottom of the seal housing to create a
restriction in the flow of the flush material through the seal chamber. This
creates enough pressure in the seal chamber to provide the required margin
above the vapor pressure of the flush liquid, which is needed to maintain
proper lubrication of the seal faces. Circumferential porting of the seal flush
liquid helps to provide uniform heat removal for the seal faces.
These pumps were re-rated/upgraded during a normal unit outage for
maintenance. The total cost of this upgrade was approximately one fifth the
cost of the installation of a set of new pumps and motors as estimated by
Omega Refinery’s Capital Project Organization. The option of hydraulically
re-rating these pumps also provided the ability to address the problem in a
timely fashion without requiring a separate unit shutdown to implement.
The re-rated pumps were placed into service in 2003. Start up of both pumps
was uneventful, and mean time between seal replacements is now close to
3 years.
Problems involving a grossly oversized pump are common to a wide variety of pump users and industries. The reasons for the pumps being oversized vary widely.
• Usually, the plants involved are old, and there are very few records
of original design operating data and maintenance history.
• Often, as in this case, a pump that the owner has available is put
into a new service without much attention being paid to operating
conditions.
• Sometimes, the conditions of service have changed, resulting in
the original pump being much too large or too small for the new
service requirements.
• Often times the system is complex, involving multiple flow paths
and/or multiple pumps, so a thorough hydraulic analysis of the system seems too daunting to conduct.
The solutions to these oversized pump problems are quite variable. In
some cases, the lowest cost and most expedient measure is to simply replace
the pump. In some situations, the existing impeller simply can be trimmed
to a smaller size, or the pump can be slowed down if the pump has a variable
speed driver. In some cases, the pump in question literally has not even been
needed and may be removed from service.
The most important lesson to be learned from this type of problem is that
any time a pump is chosen, be it for a new service or to replace a pump
© 2008 Taylor & Francis Group, LLC
278
Pump Characteristics and Applications
in an existing system, a thorough analysis of system flow and total head
requirements will insure a lower cost and more reliable pump, with the least
amount of energy consumption and maintenance costs.
B. The Case of the Unreliable Refrigerant Pump
1. Background
The Chilly Air Conditioning Company manufactures residential and commercial air conditioners, which use R-410A as the refrigerant. The Chilly
Air Conditioning Company plant has a bulk R-410A delivery system that
includes a storage tank and a pump to circulate the R-410A in a loop, which
feeds air conditioner charging stations located throughout the plant.
The pump used for this service is a horizontal multistage side-channel
pump. The side-channel pump design is similar to a regenerative turbine
(see Chapter 4, Section XIII) in that the impeller makes regenerative passes
through the liquid. The side-channel pump style includes star-shaped
impellers, with a more traditional curved vane impeller for the first stage
for lower NPSHr. The pump has similar hydraulic characteristics to a regenerative turbine pump, including generating relatively high total head and
a steeply rising head–capacity curve. Figure 9.2 shows the head–capacity
curve for the side-channel pump used in the R-410A delivery system. The
design of side-channel pumps allows pumping of liquid–gas mixtures with
1000
900
800
Total head (ft)
700
600
500
1750 rpm
400
300
200
100
0
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Flow rate (gpm)
FIGURE 9.2
Head–capacity curve for the side-channel R-410 refrigerant pump discussed in Case B.
© 2008 Taylor & Francis Group, LLC
39
40
Case Studies
279
up to 50% vapor, eliminating possible air or vapor locking problems. The use
of this type of pump for this system was necessitated by the relatively high
head required during certain operating periods, the requirement that flow
not vary widely in spite of a widely varying system head, and the fact that
R-410A is being pumped at near boiling conditions, thus containing large
amounts of vapor.
There have been problems with this pump in the Chilly Air Conditioning
Company plant almost from the beginning. The original pump was installed
with a mechanical seal, but the seal had frequent failures. This resulted in
leakage of the expensive R-410A refrigerant, plus a lowering of the reliability
of the refrigerant bulk delivery system, causing disruptions of plant operations. Through careful monitoring, better seal materials, and closer attention
to pump alignment, the 6-month average life for the seals was improved
to about a year, but this was still not acceptable to the company reliability
engineer, who has been accustomed to an interval of 3 years or more with
other pumps in the plant. The mechanically sealed pump was replaced with
a sealless magnetic drive version to eliminate the possibility of seal leakage.
(See Chapter 5, Section V.B for a discussion of magnetic drive pumps.)
The newly installed magnetic drive pump failed after only 3 months of
operation, making very high squealing sounds before seizing. With this
being the sixth failure over a 2.5-year period, the reliability engineer asked
the author to investigate the pump and the refrigerant delivery system to
determine and correct the cause of the frequent pump failures.
2. Analysis of the Problem
The pump is controlled with a fixed discharge pressure, currently set at 375
psig. The pressure at the surface of the liquid in the storage tank is expected to
be the vapor pressure of the liquid, which of course varies with temperature.
At times the actual measured pressure of the vapor space above the liquid
in the storage tank has been as high as 10 psi over the fluid’s vapor pressure.
Chilly’s Chief R&D Engineer believes that this higher than expected pressure in the tank is likely a combination of two causes. First, since the tank
receives sunshine at least part of the day, it is quite likely that the temperature of the vapor space is higher than that of the liquid below it. Since the
pressure on the surface of the liquid is determined by the temperature of
the vapor, this would account for the measured pressure in the tank being
higher than expected at times. Second, it is believed that the vapor contains
minute amounts of noncondensable gases (mostly nitrogen), which would
have the effect of increasing the vapor pressure.
Since the refrigerant storage tank is located outside, the temperature of the
R-410A liquid, as measured at the bottom of the storage tank, varies from
29°F to 87°F, depending primarily on ambient temperature. Since the pump
is being controlled to a fixed discharge pressure of 375 psig, the total head
of the pump will fluctuate widely through the year. During winter months,
© 2008 Taylor & Francis Group, LLC
280
Pump Characteristics and Applications
when the temperature in the tank is the coldest, the pump’s suction pressure would be the lowest. Therefore, the total head would need to be at the
high end of its range (corresponding to the lowest flow point), to meet the
fixed 375-psig discharge pressure. Conversely, during summer months,
the pump’s suction pressure would be the highest. At that point, the total
head would be at the low end of its range, with the pump running the furthest out on its head–capacity curve.
The investigators decided that the failures might be caused by the pump
operating outside of its normal operating range. To analyze this, the team
decided that the total head, TH, should be calculated at the extremes of operating temperatures to ascertain the resulting range of flow and compare that
with acceptable operating points on the pump curve.
For example, at 29°F, a temperature properties table shows the vapor pressure of R-410A to be 108.7 psia, or 94 psig, and specific gravity to be 1.18. The
suction pressure at the pump can only be estimated, as there is no pressure
gauge located right at the pump suction. The static head of liquid from the
low level of the storage tank is 2.5 ft above the centerline of the pump. Friction
losses in the suction line are unknown, and cannot be measured without a
gauge at the pump suction. Therefore, it is assumed that the suction pressure
at the pump is approximately equal to the pressure in the storage tank, i.e.,
the friction loss in the suction line approximately equals the static suction
head. So, at a temperature of 29°F, pump suction pressure is estimated to be
94–104 psig (i.e., ranging from vapor pressure to 10 psi over vapor pressure
as discussed above).
With a discharge pressure of 375 psig and a suction pressure of 94 psig,
the pump differential pressure would be 375 − 94 = 281 psi. Using Equation
2.7, the total head of the pump is converted from 281 psi to 550 ft. A similar exercise is performed at 104 psig suction pressure, and then at several
other temperatures in the operating temperature range, with results shown
in Table 9.1. The table then shows what the range of pump capacity would
be, based on the pump performance curve shown in Figure 9.2, for the calculated range of total head.
The calculated results shown in Table 9.1 revealed that the pump capacity
varied from a low of 31.8 gpm during the coldest months to a high of more
than 39.5 gpm (i.e., off the end of the curve) in the warmest months. With this
data, the following three potential root causes of the magnetic drive pump
failure were considered:
a. Cavitation
NPSHa at the pump is impossible to measure in the field as described in
Chapter 3 and Equation 3.9, due to the lack of a pressure gauge at the pump
suction. Furthermore, it is impossible to hear the classic crackling sounds
that usually accompany cavitation, due to the high frequency whining
noise that is typical of a side-channel pump. While the side-channel pump
impellers are able to pass liquids with high vapor content, the first stage
© 2008 Taylor & Francis Group, LLC
© 2008 Taylor & Francis Group, LLC
29
68
76
87
Ambient
Temperature
(°F)
108.7
208.7
235.7
276.7
Vapor Pressure
at Temperature
(psia)
94
194
221
262
Vapor Pressure
at Temperature
(psig)
1.18
1.09
1.07
1.03
Specific Gravity
94–104
194–204
221–231
262–272
Suction
Pressure Range
(psig)
271–281
171–181
144–154
103–113
Differential
Pressure Range
(psi)
530–550
362–383
311–333
231–253
TH Range (ft)
Total Head for the R-410 Refrigerant Pump at Several Ambient Temperatures and Resulting Pump Capacity per the Pump
Performance Curve Shown in Figure 9.2, and with Fixed Discharge Pressure of 375 Psig
TABLE 9.1
31.8–32.3
36.2–36.6
37.5–38.0
>39.5
Capacity per
Pump Curve
(gpm)
Case Studies
281
282
Pump Characteristics and Applications
impeller is still a traditional centrifugal pump impeller and therefore is
likely subject to the usual NPSH limitations, even though the pump may
pass the vapors.
The net positive suction head available would vary, depending on how
high the pressure is in the suction vessel over the vapor pressure of the
fluid. The worst case, from the standpoint of NPSH, would be if the pressure at the liquid surface is equal to the vapor pressure of the liquid. In
that case, the NPSH available would equal the static suction head (difference between liquid level in the tank and the impeller centerline) minus
the friction losses in the suction line. The friction losses in the suction
line could only be estimated at that point, because there is currently no
pressure gauge located right at the pump suction. There is a flow limiting
valve and a screen in the suction line. It is quite likely, with no margin
being allowed from the suction pressure being above vapor pressure, that
the NPSH available would be less than the NPSHr for the pump, even at
healthy operating points on the pump curve. However, the pump manufacturer has asserted that the side-channel pump should be able to pass
the vapors from cavitation without any problem. Given this assertion, and
given the likelihood that there is some pressure in the tank over the vapor
pressure, cavitation cannot be determined with certainty as the root cause
of the failures. But cavitation certainly remains a prime suspect as a contributing cause.
b. Operation Off the Right End of the Pump Performance Curve
During the warmer months, with the suction pressure at the highest level,
the pump differential pressure would be at the lowest level and flow at the
highest level, as Table 9.1 shows. Internal recirculation within the pump
likely would be higher when the pump is operating off the right end of
the curve, which could lead to excessive vibration. Although this analysis
shows that pump would likely fail due to operating too far off the right end
of the curve when the ambient temperature is 87°F, this scenario did not
likely cause the most recent failure of the mag drive pump, for the simple
reason that the ambient temperature at the time of the failure was 48°F, and
the highest temperature achieved over the 2 days prior to the failure was
68°F.
The Chilly Company has observed in the past that the magnetic drive
pump itself adds no more than 7°F to 8°F to the liquid when the ambient
temperatures are mild and the tank is not too low. Even at 76°F (the highest
observed ambient temperature in the 2 days leading up to the failure plus
an allowance of 8°F rise due to the magnetic flux in the mag drive), the furthest out on the curve that the pump would run would be at about 38 gpm,
as shown in Table 9.1, a point well on the pump curve. In conclusion, while
running too far out or off the pump curve is a concern during the warm
months of the year, this does not seem to have been the cause of this particular failure. It does, however, mean that a new control method or control point
© 2008 Taylor & Francis Group, LLC
Case Studies
283
must be established to keep the pump from running this far out on the curve
during warmer months.
c. Inadequate Circulation of Fluid through the Magnetic Drive Pump
The third potential root cause that was examined was whether the fluid
was not adequately directed to the two separate cooling/lubrication paths
that are required inside a magnetic drive can, one to cool the magnets and
the second to lubricate the bearings. From an inspection of the failed components and based on the thrust bearing manufacturer’s failure report, the
thrust bearing in the magnetic drive pump apparently failed due to dry
running. One possibility is that the flow path to the thrust bearing was
not sized to permit adequate flow in this flow path. The pump operating
with a low differential head and near or off the right end of its curve could
possibly not deliver adequate flow to lubricate the mag drive bearings.
Once again, this is not a factor that can be conclusively stated as being the
root cause of the problem, but it was deemed to be a possible contributing
factor.
3. Solutions and Lessons Learned
Based on the analysis, and with several potential contributing factors to the
problems with the pump, the following changes were recommended by the
author and were implemented by Chilly Air Conditioning Manufacturing
Company.
• A new pressure regulator was installed to reset the control pressure
to 440 psig, 65 psi higher than the original setting of 375 psig. This
additional 65 psi, when factored into the calculations that created
Table 9.1, resulted in a TH range of 377–677 ft over the range of operating temperatures, which resulted in a capacity range of 28.5–36.3 gpm,
an acceptable flow rate in all cases, and within the healthy operating
range for the pump.
• An additional 4 ft of static suction head and NPSHa was created by
moving the storage tank up four feet. Additionally, the unneeded
excess flow valve and screen were removed from the suction line,
reducing suction line friction losses and adding even more to
NPSHa.
• A pressure switch was installed to trip the pump if discharge pressure drops below 350 psig, which could indicate that the pump had
lost flow. This will protect the magnetic drive from failing catastrophically if flow is lost to the pump. A separate temperature control was installed to trip the pump in the event the magnetic drive
can exceeds a set point temperature, a redundant protection against
loss of flow.
© 2008 Taylor & Francis Group, LLC
284
Pump Characteristics and Applications
• The pump manufacturer was asked to make several modifications
to the pump, including opening the port that directs flow to lubricate the thrust bearing and changing the material combination of
the bushings.
• A pressure gauge was installed at the pump suction flange.
This combination of actions has resulted in 6 years of continuous operation
of this pump, with no failures. The major lessons that were learned from this
experience were to look carefully at the operating range of the pump during
system fluctuations and to address the problem from several directions if
there is not one clear root cause of a failure.
C. The Case of the Vibrating Vertical Turbine Pump
1. Background
The Mighty Steel Corporation has a component cooling water system that
has a widely varying load, depending on process and building cooling
needs. Part of the cooling system includes a large cooling tower with a vertical lineshaft turbine style pump (see Chapter 4, Section XI and Figure 4.34),
mounted over the cooling tower basin. Due to the wide heat load variation
in this system, the flow of the cooling tower pump has been controlled by
a flow control valve. This system has become quite old and unreliable, and
engineering management has decided to modify this system to remove the
flow control valve and to vary the flow by varying the pump speed.
The difference in elevation between the liquid level in the cooling tower
basin and the cooling tower inlet is minimal, thus, this system has a minimal
amount of static head.
Since the system head for the cooling water system is primarily composed
of friction head, this type of system is ideally suited for controlling flow with
a variable speed pump. This type of operation will keep the pump running
near its BEP virtually all the time, as well as minimizing energy consumption and will eliminate the need for the flow control valve with its attendant maintenance. The engineering team prepared to convert the cooling
tower pump to a variable speed pump by running the motor through a variable frequency drive (VFD). Variable speed pumping systems and VFDs are
described in more detail in Chapter 6, Section IV.
Mighty Steel’s Electrical Engineer reviews the electrical requirements for
the VFD and confirms that the premium efficiency motor presently being
used on the cooling tower pump is adequate for running through a VFD. The
purchase specification is prepared for the VFD, it is purchased and installed,
and the pump is started up. The pump at first performs well, both running
at full speed and at several reduced speeds. Suddenly, on the second week of
operation, the pump begins to run very loudly and to exhibit excessive vibration. The operators quickly shut the pump down, and the utility engineer
© 2008 Taylor & Francis Group, LLC
Case Studies
285
and the maintenance supervisor meet to figure out what is happening with
the system.
2. Analysis of the Problem
With the cooling tower pump shut down, the pump’s shaft runout at the
mechanical seal was checked with a dial indicator mounted on the pump
discharge head. The runout was determined to be 0.003 in, which is within
the manufacturer’s specification. The pump turns by hand, with no audible
suggestion of any foreign material in the pump and no feeling of rubbing in
the bearings. The motor was briefly bump-started to confirm correct rotation. The cooling tower basin was checked thoroughly for foreign materials
that might have become lodged in the screen at the suction bell of the pump,
but the basin appeared clear with no significant debris.
The next step in troubleshooting was to run the pump to take vibration
readings, including vibration frequency, which the utility engineer hoped
would help determine the cause of the vibration. The engineer also checked
for surface vortices that might have been drawing air into the pump. However,
when the pump was run at full speed, there was no observed vibration or
noise. This immediately threw suspicion on the newly installed VFD.
The pump was run through a series of reduced speed tests, with each test
being done at 100 rpm slower speed than the preceding test. For the first five
reduced speed tests, the pump ran fine, with no increase in noise or vibration.
At a speed of 1200 rpm, the vibration and noise of the pump were noticeably
higher. When the pump was run at 1100 rpm, the noise and vibration were
too high to allow the test to continue, and the pump was shut down. After
these test results were reviewed with the pump manufacturer, the utility engineer felt quite certain that the pump was experiencing a resonance problem.
(Resonance is discussed in Chapter 4, Section XI.) Briefly, resonance occurs
when a pump operates at a running speed that is too close to its natural frequency. This is most likely to occur with relatively long flexible pumps, such as
vertical turbine pumps, because the natural frequencies of most other pumps
are well above their operating speeds. It is quite common for vertical turbine
pumps to pass through the first natural frequency at start up, on their way up
to full running speed. Generally, these types of pumps are designed by the
manufacturer to operate at a speed that is at least 25% away from the natural
frequency. This was not a problem when the pump was running at a fixed
speed of 1780 rpm. But when controlled by the VFD, the pump was apparently
allowed to run at a speed that was too close to the natural frequency.
Natural frequency on a vertical turbine pump is not easy to calculate accurately, as it is a factor of many variables (e.g., column length and diameter,
weight of the bowl assembly, design of the discharge head, weight and natural frequency of the motor, etc.). Some factors, such as how the pump is
mounted, supported, and piped also affect the natural frequency and are out
of the control of the pump manufacturer.
© 2008 Taylor & Francis Group, LLC
286
Pump Characteristics and Applications
The utility engineer reviewed the operational logs for the week during
which the pump had run fine on the VFD when it was first started up and confirmed that the operating speed of the pump had always exceeded 1200 rpm
during that week. The morning of the first high vibration event, the load
on the cooling water system had dropped as some machinery went off line,
resulting in the speed of the pump dropping below 1200 rpm. Based upon the
testing, it seemed that the natural frequency of the installed pump was close
to 1100 cycles per minute, since the pump vibrated the most at that speed.
The VFD was set at 1000 rpm and the pump was started again. Vibration
levels were slightly above normal, but not nearly as high as at 1100 rpm. At
900 rpm, the pump operated smoothly again. The utility engineer continued
testing down to 800 rpm, the slowest expected operating speed of the pump,
with no further high vibration level. Checking at 50-rpm increments rather
than at 100-rpm increments revealed that an increase in vibration occurred
between speeds of 950 and 1250 rpm, with a maximum vibration level occurring at or near 1100 rpm.
3. Solutions and Lessons Learned
Fortunately, the VFD was able to be programmed to not dwell in specified speed ranges, but rather to pass quickly through them. So, it was programmed to avoid the range of 950 to 1250 rpm. If the need for less flow was
called for by the temperature indicator input signal for the VFD, the drive
would not reduce speed below 1250 rpm unless the temperature signaled
the motor speed to drop below 950 rpm. This resulted in situations where the
pump supplied more cooling water than needed, but this kept the pump out
of the speed zone where it was operating in resonance.
The author has seen other resonance problems with vertical turbine
pumps running with VFDs, especially ones such as this case, where the
pump initially was purchased as a constant speed pump and then subsequently converted to running with a VFD. With constant speed pumps (ones
that have no VFD), resonance problems are considerably more rare, but still
occasionally crop up. This may be because the type of pump support or piping significantly changes the natural frequency of the installed pump, or in
some cases, there may have been a miscommunication of natural frequency
information about the motor.
With fixed speed pumps, the solution is not as easy. Rather, the solution
with a fixed speed pump in resonance is to change the natural frequency
of the installed pump. There are many techniques that may be employed
to accomplish this. These may involve reducing the natural frequency (i.e.,
making the pump more flexible) by, for example, opening up the coupling
window in the motor support. Another solution that has worked to lower the
frequency is to put 0.100 in thick washers on the motor support face where
the pump and motor come together. Since the male register on the motor
support is typically at least 0.125 in thick, there is still a register to mate with
© 2008 Taylor & Francis Group, LLC
Case Studies
287
the female fit on the motor, but the resulting gap between pump and motor
lowers the overall natural frequency.
Or the fix may be to raise the natural frequency of the pump (i.e., making
the pump assembly stiffer) by welding ribs on the discharge head and motor
support, changing the motor support to a larger diameter, etc.
In one field resonance problem, the solution was to change the natural
frequency by changing the mass, not the spring constant, of the assembly.
In that situation, the mass was changed by bolting additional weight in the
form of multiple square pieces of 1 in steel plate on the sides of the motor.
D. The Case of Too Many Pumps
1. Background
Often when working on pump systems, the mentality is that one can never
have “too much pump.” The first effort in solving a problem is often just to
add more pumping capacity. This case highlights how this mentality can
lead to unnecessary capital expense and higher operating and maintenance
costs. It also shows that the best way to identify problems within a system
and to make corrections is to first have a clear picture of how the system
operates, which in this case is aided by the use of commercial piping analysis
software.
The Delta Power Company has an auxiliary cooling water system as part
of their Uppa Creek co-generation electric power plant. The auxiliary cooling water system provides cooling water to equipment that is essential for
the generation of electricity, such as lube oil coolers, fan cooling coils, air
compressors, and generator coolers, along with a variety of other equipment
needed for plant operation. Auxiliary cooling water systems are found in
virtually every type of commercial power generating plant, as well as in
many other types of process plants.
This case describes a scenario that often occurs in auxiliary cooling water
systems. The system requirements change over time as the cooling requirements of the plant change and/or as the components in the piping system
wear and corrode. This sometimes results in the need to rebalance the system, modify existing pumps, or add additional pumps to meet the new conditions in the cooling water system.
When the Uppa Creek power plant was initially designed, the auxiliary
cooling water requirements for the various loads were determined, and the
cooling water system was designed to meet the system’s cooling loads. Each
pipeline in the system was sized to achieve the desired flow rate, and the
auxiliary cooling water pumps and motors were selected and sized to meet
the total system flow requirement. Balancing valves were inserted into each
circuit to limit the flow to each individual load to the set value for that load.
At the time the system was started, the auxiliary cooling water system provided sufficient cooling water load for full plant operation with only two
© 2008 Taylor & Francis Group, LLC
288
Pump Characteristics and Applications
auxiliary cooling pumps operating, in parallel. A third identical pump was
installed as a backup for the two operating pumps.
Over time, new equipment had been added as the plant’s power generating capacity was increased. The additional cooling load requirements for the
new components were added to the auxiliary cooling water system loads.
Auxiliary cooling water systems are often incorrectly considered by plant
operating personnel to be infinite resources. As a result of this incorrect
assumption, the new cooling loads were added to the auxiliary cooling water
system without regard to the effects on the existing loads.
After years of operation, the equipment wore down and heat exchangers
became fouled and partially plugged, causing a reduction of flow through
the system. This reduction in flow in the circuits caused the equipment to
overheat. To correct this problem, the plant operators opened balancing
valves on the individual circuits to increase the flow rate to the components
being cooled.
After still more time of continuous operation, one of the two operating auxiliary cooling water pumps tripped due to an overloaded motor. The plant
operators immediately started the standby auxiliary cooling water pump to
return the system to normal.
2. Analysis of the Problem
After investigating the problem, it was determined that the motor tripped
due to excess current. It was further determined that the two running pumps
were operating off the end of their pump curves, indicating that the motors
were overloaded. The third auxiliary cooling water pump was then started,
resulting in all three of the pumps running with no pump on standby. The
flow rate through each of the three cooling water pumps was within the
manufacturer’s recommended range of flow, and the motors were no longer
overloaded.
It appeared that the solution was to operate three auxiliary cooling water
pumps at all times to meet the increased needs of the cooling water system.
However, the plant’s operating procedure requires a standby pump for all
critical piping systems. The auxiliary cooling water system is a critical system, so a design change was immediately initiated to install a fourth auxiliary cooling water pump. The utility’s project engineering group was tasked
with implementing this change. Since this modification to the auxiliary cooling water system is required to increase plant reliability, the project was fast
tracked. The scope of work involved specifying the fourth auxiliary cooling
water pump, motor, and switchgear. In addition, design changes needed to
be made to connect the new pump to the existing suction and discharge
manifolds, design the pump foundation, and add the instrumentation, pipe
supports, electrical conduit, and cabling.
The first step involved specifying the new pump, since that step represented the longest lead time. The original idea was to install a fourth auxiliary
© 2008 Taylor & Francis Group, LLC
289
Case Studies
Building 255
FCV-31
FCV-28
FCV-25
FCV-22
HX 30
HX 27
HX 24
HX 21
FCV-38
HX 37
FCV-41
HX 40
FCV-44
HX 43
FCV-47
HX 46
FCV-54
HX 53
FCV-57
Building 123
HX 56
FCV-60
HX 59
HX 62
FCV-63
Cooling tower
ACCW P17
ACCW P15
CT basin
ACCW P16
cooling water pump identical to the existing three pumps. Since additional
cooling loads were added to the cooling water system over the years, project
engineering wanted to ensure that the existing three pumps could still meet
the current plant cooling loads. A system audit was conducted on all the
loads being fed by the auxiliary cooling water system. It was discovered that
even with the added system loads, the two existing pumps had sufficient
capacity to meet the flow needs of the auxiliary cooling water system. As a
result of the system audit, it was determined that a full hydraulic analysis
of the auxiliary cooling water system was necessary to determine why the
current cooling water system needs were unable to be met with two pumps.
The project engineering team decided to model the auxiliary cooling water
system using commercially available fluid piping software. Figure 9.3 shows
the piping system model.
After the piping system model was created, the next step was to set all
the loads for the heat exchangers to their design flow rates. A full hydraulic
network simulation was then performed. While reviewing the results, it was
confirmed that the two auxiliary cooling water pumps are well within the
normal operating region of their curves and should be able to provide sufficient flow for the system (see Figure 9.4).
An additional system hydraulic analysis was performed of the auxiliary
cooling water system with all three pumps running (the way the system
Building 51
FIGURE 9.3
This is the piping schematic of the auxiliary cooling water system using piping simulation
software. (Courtesy of Engineered Software, Inc., Olympia, WA.)
© 2008 Taylor & Francis Group, LLC
290
Pump Characteristics and Applications
300 16.5 in
60
275
70
75
80
83
250
83
225
Total head (ft)
200
175
Three-pump
operation
13.25 in
Two-pump
operation
150 12 in
125
60
70
75
100
80
81.8
80
80
75
50
25
0
100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100
Flow rate (U.S. gpm)
FIGURE 9.4
The pump curve shows where the pump should be operating with both two and three auxiliary cooling water pumps in operation. (Courtesy of Engineered Software, Inc., Olympia, WA.)
currently was operating). The flow rates through these pumps were well to
the left of each pump’s best efficiency point (BEP) (see Figure 9.4). When these
calculated results were reviewed, it was also discovered that there were large
differential pressures across the balancing valves with the three pumps in
operation.
Next, it was time to compare the results of the piping system model to the
operating data of the actual auxiliary cooling water system. With all three
pumps running, the observed discharge pressures for the pumps were lower
than the calculated discharge pressures as predicted by the hydraulic simulation. This indicated that the actual pumps may have been running further
out on their pump curves than the hydraulic model had predicted, indicating greater flow through the pumps and system.
In looking at the positions of the balancing valves serving the auxiliary
cooling loads, it was discovered that the majority of the valves in the actual
system were in their fully open position. As a result, the system was not
balanced and the flows were much greater than their design values. Further
investigation revealed that after each new load had been added to the auxiliary cooling water system, the system was not rebalanced. The plant operators said that over time the various heat exchangers required additional flow
from the auxiliary cooling water system to keep the outlet temperatures of
the loads at their required values. They responded by opening the balance
© 2008 Taylor & Francis Group, LLC
291
Case Studies
valves to increase the flow, which also returned the operating temperature
to the required value. This continued for many months until it seemed that
all the balancing valves in the auxiliary cooling water system were at their
fully opened position.
A hydraulic analysis of the system was performed with all balancing
valves in their fully opened position, the way the auxiliary cooling water
system was currently being operated. It was discovered that with all balancing valves fully opened and two auxiliary cooling water pumps in operation, the pumps were running off the end of their pump curve (Figure 9.5).
This caused a large current draw on the motors driving the auxiliary cooling
water pumps. This was the reason that the pump tripped on high current,
requiring the third pump to be placed in operation in order for the system to
operate without tripping a motor.
Another analysis of the auxiliary cooling water system was performed,
this time with three pumps in operation and all balancing valves open.
The calculated discharge pressure for the auxiliary cooling water pumps
closely matched the observed discharge pressure in the system. In addition,
the pressures through the remainder of the system closely matched the calculated values of the hydraulic analysis. This indicated that the hydraulic
model of the auxiliary cooling water system closely matched the operation
of the actual system.
300 16.5 in
60
280
70
75
260
80
83
240
83
Total head (ft)
220
200
180
160
140
120
13.25 in
80
12 in
60
70
75
100
81.8 Three-pump
operation
80
80
80
Two-pump
operation
60
40
100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100
Flow rate (U.S. gpm)
FIGURE 9.5
With two pumps operating and all balancing valves open, the flow rate through the auxiliary
cooling water pumps ran off the end of their pump curve, causing the motors to be overloaded.
(Courtesy of Engineered Software, Inc., Olympia, WA.)
© 2008 Taylor & Francis Group, LLC
292
Pump Characteristics and Applications
3. Solutions and Lessons Learned
In looking at the calculated results with all throttle valves in their 100% open
position, the analysis showed that flow rates to all the loads in the auxiliary
cooling water system greatly exceeded the required flows as determined in
the audit. As a result, it was decided that the first step in correcting the problems with the auxiliary cooling water system was to balance the system by
adjusting the balancing valves to limit all system loads to their design values.
The piping analysis software was used to determine the valve positions
for each balancing valve in the system. This was accomplished by entering
the manufacturer’s Cv values for the balancing valves into the piping system
model. The hydraulic analysis of the system was conducted with two pumps
in operation (the original normal system operating condition). The program
calculated the position of each of the balancing valves in the system. Once
this was accomplished, there was a starting point for balancing the system
with two pumps in operation.
So that the system could be balanced with the plant in operation, each
balancing valve was placed in the position corresponding to the normal
operating condition of two pumps in service. Since the system was currently
running with all three auxiliary cooling water pumps, each of the loads
would be assured of getting more than its required flow during the changing of the system.
To ensure that the piping system model and the actual piping system were
still in agreement, a hydraulic analysis was performed on the three auxiliary
cooling pumps with the balancing valves set up for two-pump operation. An
ultrasonic flow meter was used to compare the actual values in the system
with the calculated values in the piping system model. The flow rates in each of
the circuits closely matched the values calculated by the piping system model.
The final step was to shut down the third auxiliary cooling water pump
and return the system to its normal operation. After the standby pump was
shut down, the flow rate feeding each of the loads was again checked with
the ultrasonic flow meter. After reviewing the results of all the flow rate measurements, it was determined that all but two of the loads in the auxiliary
cooling water system were receiving their design flow rates.
An investigation of the two heat exchangers was conducted. Pressure
gauges were placed across the heat exchangers in question, and it was discovered that the observed differential pressure was higher than what was
indicated in the piping system model. The heat exchanger manufacturer was
contacted and was able to supply a curve showing the differential pressure
across the heat exchangers for a range of flow rates. The piping system model
for the two heat exchangers in question was compared with the data supplied by the manufacturer. The model was updated with the manufacturer’s
supplied information, and it was determined that the two heat exchangers in question still had a high differential pressure for the observed flow
rate. The higher differential pressure observed across each heat exchanger
© 2008 Taylor & Francis Group, LLC
Case Studies
293
suggested that the water sides of the heat exchangers were fouled or partially
blocked. The manufacturer suggested performing a heat balance for each
heat exchanger to confirm the assumption.
The inlet and outlet temperatures for both the load side and the auxiliary
cooling water system side were recorded, along with the observed flow rates.
The results confirmed that the heat transfer rates in the exchangers were
below their design values. This information, along with the abnormally high
differential pressure across the auxiliary cooling water tube sheets, was a
clear indication of fouling in the heat exchangers in question. As a result of
this information, the heat exchanger tube sheets were cleaned during the
next plant shutdown.
Often when a problem is encountered in piping systems (in this case, the
auxiliary cooling water pump motor tripping on high current), the first
response is to “power through the problem” by adding another pump. This
results in masking the problem instead of correcting it. In this example, with
a clearer picture of the system provided by the software analysis, the plant
staff was able to isolate the real problem, try a variety of options, and correct
the problem without having to resort to purchasing and installing additional
equipment.
Once it was shown that the hydraulic analysis software produced an accurate model of the auxiliary cooling water system, all the testing and recommendations were able to be checked with the model. This eliminated the
need to take the auxiliary cooling water system off line (along with the entire
plant) to test the effectiveness of the proposed changes. As it turned out, the
entire balancing of the system was able to be performed without affecting
the operation of the generating station in any way. The only time any operational changes were required was when the two fouled heat exchangers
were removed from service for cleaning, and this was accomplished during
a scheduled plant outage.
By balancing the system, the cost was minimal, consisting only of the time
needed to perform the analysis and rebalance of the system, something that
should be done after every system modification.
The option of adding a fourth auxiliary cooling water pump would have
resulted in a major capital cost for the pump, motor, switchgear, and interconnecting piping. This change would also have required a construction
outage to tie in the new auxiliary cooling water pump. Finally, there would
have been the added cost of continually running and maintaining the third
auxiliary cooling water pump.
E. The Case of Too Few Pumps
1. Background
The Pure Pharmaceutical Company has a large design and manufacturing
campus on the east coast of the United States, with production buildings
© 2008 Taylor & Francis Group, LLC
294
Pump Characteristics and Applications
spread over a large area. The complex has a central utility plant that includes
a purified water system that provides purified water to users in six different production buildings located around the complex. Hygienic standards in
the pharmaceutical manufacturing industry do not permit making purified
water at the point of use, to have the purified water manufacturing more
carefully controlled. Furthermore, purified water is required to continuously
circulate at a minimum velocity of roughly 3 ft/s to maintain turbulent flow
to avoid possible contamination. The purified water system was designed as
a loop around the manufacturing complex, as shown in Figure 9.6.
When the purified water system was originally designed, the ultimate
requirements for purified water at the various production buildings were
estimated by the Pure Pharmaceutical process engineers. Any or all of the
cells could be idle at any one time. Thus, the process engineers were able to
establish a range of flow requirements for the pumps in the purified water
circulation loop. Even if all of the production cells are using their estimated
flow rate, flow has to continue to circulate though the remainder of the loop
at the minimum 3-ft/s velocity required to maintain turbulent flow. The
resulting required flow rate for the pumps varies from 110 gpm (when none
of the production cells are using purified water) to 210 gpm (when all of the
production cells are using purified water). The pumps are controlled via a
pressure control valve located near the end of the loop.
Because the purified water loop covers a very large complex, with over
5300 ft of 2 1/2 in and 3 in piping in the loop, a significant amounts of
dynamic friction head was present in the system. When the system head
losses calculations were completed, total system head (consisting entirely of
dynamic friction head) was computed to be 536 ft at 210 gpm.
Makeup water from
reverse osmosis system
Pressure control valve
Level control valve
Return pump
Supply pump
Purified water ports at production cells
FIGURE 9.6
Flow diagram of the purified water system described in Case E.
© 2008 Taylor & Francis Group, LLC
295
Case Studies
The design engineer had considered having only a single pump in the system, but that would have meant a total head for that pump of 536 ft. This
would have resulted in local pressures in the system well over 200 psi, which
was the upper pressure limit for a number of the sanitary diaphragm valves
in the system. (For a specific gravity of 1.0, 536 ft is equivalent to 232 psi, per
Equation 2.7.) By distributing the system head between two pumps, a supply
pump and a return pump, the maximum pressure that the system could see
at any point remained below 200 psi.
The desired physical location of the return pump is close to the end of
the piping system, so most of the head loss in the system occurs prior to the
return pump. Once the pumps were located in the system, the design conditions that were set for the two pumps were as follows:
• Supply pump – flow of 210 gpm at TH of 424 ft
• Return pump – flow of 210 gpm at TH of 112 ft
In addition, both pumps need to be checked for operability at their lowest
anticipated flow rate of 110 gpm. A pressure switch was installed in the system with a set point to trip the return pump if the system were inadvertently
blocked downstream of the pump.
Figure 9.7 shows the performance curve for the return pump that was chosen. For the supply pump, the only way to get the necessary higher head
150 145 mm 20
30
40
50
140 140 mm
60
130
63.6
120
60
110
100
Head (ft)
90
80 115 mm
20 30
70
60
50
40
50
40
40
40
30
20
10
0
20
40
60
80
100
120
140 160 180
U.S. gpm
200
220
240
260
FIGURE 9.7
Performance curve for the return pump for the purified water system in Case E.
© 2008 Taylor & Francis Group, LLC
280
300
296
Pump Characteristics and Applications
was to choose the pump shown in Figure 9.8. This pump was capable of
much higher flow than the system called for, as the BEP of the supply pump
was nearly 600 gpm. The engineer would have liked to have found a pump
with a smaller impeller and casing than the pump shown in Figure 9.8, so
that the BEP flow would be closer to the design flow. However, sanitary
guidelines required that this pump be a hygienic overhung end suction style
pump. This pump style includes many particular features that are specific
to hygienic services, such as an open impeller, highly polished 316-L stainless steel wetted parts, FDA-approved elastomers, and clean-in-place (CIP) or
steam-in-place (SIP) provisions.
There are a limited number of suppliers and hydraulic options available
for this special pump design, none of which have the low specific speed that
would give low flow and high head in a single small impeller. The only way
to make the high head that was needed for the supply pump was to use
the large impeller found in the pump in Figure 9.8. Even though the pump
was running fairly far back on the curve, the engineer reasoned that it was
within the range that the manufacturer said was acceptable. The engineer
was satisfied that the pump would work well at 210 gpm, as this was well
within the operating window of the pump shown on the pump performance
curve. Furthermore, the pump manufacturer had shown the minimum flow
450 250 mm
425
40 45
50
245 mm
55
60
65
70
400
72.4
71.8
375
Head (ft)
350
325
300
275
250
195 mm
40 45 50 55
60
65
225
70
71.8
200
175
50
100
150
200
250
300
350 400
U.S. gpm
450
500
550
600
650
700
FIGURE 9.8
Performance curve for the original supply pump for the purified water system in Case E. This
pump is capable of much higher flow than needed for the system.
© 2008 Taylor & Francis Group, LLC
Case Studies
297
of the pump to be 65 gpm, which was well below the anticipated lowest flow
of the pump at 110 gpm.
Not long after the purified water system was started up, the supply pump
began to experience excessive noise and vibration. This ultimately resulted
in the pump seizing. When the pump was pulled from service, the impeller
was found to have rubbed against both the front and back sides of the casing,
and several vane tips had broken off the impeller. The pump wet end was
replaced, and less than a day later experienced a similar failure. The reliability engineer was assigned to study the problem.
2. Analysis of the Problem
A review of the supply pump curve showed that the pump was running
between 18% and 35% of the BEP when it ran in the range of 110–210 gpm.
A review of the production flow data showed that the pump had indeed
been operating within this range. In fact, several production buildings had
been idle for the past few months, so the flow rates were typically below
150 gpm at the times that the pump failed. The reliability engineer immediately became suspicious that this was the root cause of the problem. She
understood the reason why this pump was chosen, but thought there might
be better, though likely more expensive, options.
Before jumping to a rash conclusion, the reliability engineer decided to
study the problem carefully to rule out other causes. She first verified motor
rotation to be correct. She next verified that there was adequate NPSH margin for the pump. She carefully inspected the installation, which looked fine.
The pump was close coupled, so there were no alignment concerns. The
motor shaft was checked for runout, and the registered fits on the C-Face
motor and pump casing were checked for concentricity and parallelism.
Nothing seemed amiss there.
The reliability engineer discussed the application further with the pump
manufacturer, and came to the conclusion that the supply pump was operating at too low a flow, which resulted in considerable internal recirculation
within the pump, as well as both higher radial and axial bearing loads than
the pump was designed to handle for continuous operation.
3. Solutions and Lessons Learned
The reliability engineer concluded that the original pump chosen for the
supply pump was not a good choice, because it resulted in operation too low
below the best efficiency point of the pump, in this case 18%–35% of BEP.
The reliability engineer proposed to split the total head required for the
supply pump into two smaller pumps operating in series. The supply pump
was pulled from service and replaced with two new supply pumps as shown
in Figure 9.9, with each of these new supply pumps designed for 210 gpm at
212 ft TH.
© 2008 Taylor & Francis Group, LLC
298
300
Pump Characteristics and Applications
30
205 mm
40
50
60
65
280
65
60
260
240
50
180 mm
66.9
Head (ft)
220
200
170 mm
30
40
50
60
65
180
40
65
160
60
40
140
50
120
100
25
50
75
100
125
150
175 200
U.S. gpm
225
250
275
300
325
350
375
400
FIGURE 9.9
Performance curve for the new supply pump for the purified water system in Case E. Two of
these pumps, operating in series, replaced the pump shown in Figure 9.8.
There was some reluctance on the part of Pure Pharmaceutical Company
management to make this change, due to the cost required to retrofit the
installation to fit the two pumps in series in the piping system, and due to
the higher cost of the two new supply pumps, compared with the original
supply pump. But, with the new supply pumps operating at 50%–100%
of their BEP when running in the 110 to 210 gpm range, the reliability
engineer figured, correctly, that the new supply pumps would operate
more smoothly and would reduce maintenance headaches. Furthermore,
the new supply pumps were operating at higher efficiency than the original supply pump, both at 210 and at 110 gpm, so overall less energy was
consumed.
One lesson learned from this experience is that just because a duty point
fits within a manufacturer’s selection window does not make it a good choice
for continuous service. Many manufacturers show operating windows for
pump selection that are broader than the pump should be run on a continuous basis (i.e., more than 2 h out of 24). As a rule of thumb, some engineers
avoid operating pumps in a wider range than 70%–120% of the BEP flow,
except for very intermittent situations. This range gets even tighter with
larger pumps and pumps with higher suction specific speeds.
Another lesson learned is that sometimes the special circumstances of the
application limit the number of pump suppliers, or the type of pump to use,
© 2008 Taylor & Francis Group, LLC
Case Studies
299
which may lead to selection of a pump operating in a nonoptimal part of the
pump curve. The author has seen this situation occur in situations where
an ANSI chemical process pump (see Chapter 4, Section XIV.B) is required,
and where the requirement is low flow and high head, causing engineers to
choose a large impeller pump (one capable of much higher flow) running well
back on the curve. The need for this situation has largely been eliminated in
the case of ANSI pumps, with the low-flow ANSI pump option that is ANSI
dimensioned but has a Barske-type impeller that produces lower flow and
higher head than a typical ANSI pump design (see Chapter 7, Section VII for
a description of Barske impellers).
Another lesson learned in this case is that sometimes it may be better to
split the head in a system, or a portion of a system between two pumps in
series so that a pump can be chosen that runs closer to its BEP.
F. The Case of the Underperforming Pump
1. Background
The Acme Widget Manufacturing Company has two makeup water pumps
that supply raw water to the plant from the nearby Lazy River. This water
is treated at the plant and used for cooling, rinsing, and diluting functions
in the manufacturing process. Normally one pump is used, but there is a
100% backup spare, since makeup water is essential for plant operations.
State environmental regulations require that the wastewater recovered from
the process be pretreated by Acme Manufacturing before discharging into
the municipal waste system for further treatment. Thus, the makeup water
pumps operate a significant number of hours per year. They are set up on an
alternator control to spread the operating hours between the two pumps, so
the two pumps both accumulate significant operating hours. The pumps are
between bearings horizontal split case pumps with a double suction impeller, similar to the pumps shown in Figures 4.16 and 4.17.
The Acme plant is more than 40 years old, and the plant operating rate
has grown and shrunk a number of times during that period, in response to
market demands, the formulation of Acme’s manufactured products, and the
aging and replacement of certain process equipment. The original makeup
water pumps are still being used, though the original motors were replaced
about 15 years ago after being struck by lightning. The replacement motors
were taken from the company’s own warehouse, salvaged from another system that had been decommissioned. The replacement motors were used even
though they were larger than needed for the makeup water pump application. (The replacement motors are 200 HP, while the original motors were
only 125 HP.) The salvaged motors were available and in good shape and
were used to avoid the capital cost of new motors. Acme Company’s electrical engineer looked at the larger motor efficiency curve, and he confirmed
that the larger motors would maintain their peak efficiency all the way down
© 2008 Taylor & Francis Group, LLC
300
Pump Characteristics and Applications
to 50% of full load. (This characteristic of many electric motors is shown on
the typical motor test data shown in Figure 4.54.) Therefore, the replacement
motors should not see a significant power loss due to reduced motor efficiency, even running at 50%–60% of full load.
It was midwinter, the time when Acme’s production rate was at its annual
low point. With the plant running smoothly, the reliability engineer decided
to run pump performance tests on some of the larger pumps in the plant.
The plant had an ultrasonic flow meter and a kilowatt transducer that the
reliability engineer used to conduct field performance tests on the pumps.
For the total head measurement on the pumps, the suction and discharge
of all of the major pumps had been fitted with pressure gauges. The normal performance test procedure called for the reliability engineer to measure total pump head (TH) using the pressure gauges and Equation 3.6, to
measure flow rate (Q) using the ultrasonic flow meter, and to measure the
KW into the motor using a kilowatt transducer. The reliability engineer
would then convert the motor input KW into HP using the conversion rate
of 1.341 per Appendix B (kilowatts × 1.341 = HP) and then multiply that
by the motor efficiency to get the motor output horsepower, which is the
same as the pump BHP. The measured values of Q, TH, and BHP were then
substituted into Equation 2.16 (using 1.0 specific gravity) to calculate the
pump efficiency at the measured flow point. It was the reliability engineer’s
practice to periodically do this field test with any pumps larger than 100
horsepower, so that he could compare the measured pump efficiency with
the efficiency shown on the original factory pump test curve. This helped
to prioritize pumps for repair. The reliability engineer figured that if, for
example, a 200-HP pump was losing 10% of efficiency due to increased wear
ring clearances, this was an effective loss of about 20 HP. With a power cost
of $0.10/kW-h and a pump that ran 3000 h/year, that extra 20 HP of wasted
power converted to an annual cost of
20 HP × 0.746 kW/HP × $0.10/kW-h × 3000 h/year = $4476/year
The prospect of gaining back this expected energy cost by restoring wearring clearances during a pump overhaul helped the reliability engineer to
justify repair work on pumps in certain situations.
When he did the field performance test on the makeup water pumps as
described above, the reliability engineer found some confusing data. The
measured flow rate and total head for one of the pumps (Pump A) were
below what he expected and seemed to indicate that the pump was not
operating on the expected head–capacity curve. Furthermore, the tested efficiency of the pump was nearly 25 percentage points below the efficiency that
he expected to measure. He switched over to the backup pump (Pump B),
and tested that pump, using the same procedures described above. Pump B
appeared to be operating on its performance curve as expected and with the
expected power draw.
© 2008 Taylor & Francis Group, LLC
Case Studies
301
2. Analysis of the Problem
The reliability engineer’s first thought was that the Pump A was badly worn
and in need of repair to renew the pump’s wear ring clearances. Examining
maintenance records for Pump A, he was surprised to find that it had just
been repaired less than 3 months previously. Given the normal quality of the
makeup water being pumped, the Pump A wear ring clearances should not
have opened up significantly in that relatively short amount of time.
As per his typical troubleshooting protocol, the reliability engineer took
vibration and bearing temperature readings on Pump A while it was operating and found that both exceeded normal expected values. This surprised
him a bit, as a pump running with increased wear ring clearances (his first
guess as to the root cause) would normally not vibrate excessively, provided
that the pump was operating in the stable operating range. His next step was
to shut down Pump A and put Pump B into service. This allowed Pump A to
be bump started, to observe the pump rotation to verify it was correct. If the
motor electrical leads on the three-phase Pump A motor had been reversed
from their correct position when the pump was reinstalled following the
repair 3 months previously, the motor and pump would have been rotating backward, which could have resulted in the symptoms being exhibited.
However, the rotation check verified that indeed Pump A was rotating in the
correct direction.
The reliability engineer carefully checked out the rest of the system to
make sure that there was not other obvious problem with incorrectly positioned valves or a partially blocked suction that might have contributed to
the Pump A symptoms. Everything checked out, so the reliability engineer
reluctantly came to the conclusion that he was going to have to take Pump
A out of service and return the rotating element to the shop. He concluded
that either the pump’s wear ring clearances had been improperly machined
during the overhaul, the ring clearances had worn prematurely due to the
introduction of foreign material into the makeup water system, or there was
some other internal problem within the pump.
When the rotating element of the pump reached the shop, the reliability
engineer was called by the maintenance foreman to the receiving area. There
the wear ring clearances of the double suction impeller were checked. The
clearances were in as-new condition, ruling out excessive wear ring clearances as the culprit. The rest of the rotating element of the pump also looked
in as-new condition. The reliability engineer next pulled all of the shop
drawings and data sheets they had from the manufacturer to see if he could
figure out what was causing this pump’s poor performance problems.
3. Solutions and Lessons Learned
Just when he was about to throw up his hands in despair, the reliability engineer discovered the problem. From the manufacturer’s pump data sheet, he
© 2008 Taylor & Francis Group, LLC
302
Pump Characteristics and Applications
saw that the pump was supposed to be operating clockwise when viewed
from the coupling end. However, as he observed the impeller mounted on
its shaft, it was obvious that the impeller would need to be operating counterclockwise when viewed from the coupling end in order for the curvature of the
impeller vanes to be correct (see discussion of this in Chapter 1, Section V). But,
how could this be, as the reliability engineer had observed that the pump
was operating clockwise when viewed from the coupling end when he had
bump-started the motor?
Suddenly, the reliability engineer understood. When the pump was reassembled following the overhaul, the impeller was inadvertently mounted
backward on the shaft. This double suction impeller has a straight bore, like
most double suction impellers. The impeller slides to the center of the shaft,
where it is keyed to the shaft and held in place by threaded sleeves. However,
it is possible for the impeller to fit on the shaft with the vanes oriented either
way. Also, it is possible for the rotating assembly to fit inside the pump casing, and for the casing cover to be closed, even with the impeller incorrectly
mounted backward from what it should be. Again, this is true for many split
case double suction pumps, a situation that is unique to this particular style
of pump and impeller. The impeller should have been checked for proper
orientation when the rotating assembly was put together after the overhaul,
but this was missed.
Thus, with a reversed impeller, the pump was acting somewhat like a
pump operating with a reversed rotation, even thought it was rotating correctly. Ordinarily, the resulting very inefficient operation would have been
noticed at startup, because of the much higher power draw that the inefficient setup produced. However, because the salvaged motors that had been
installed on these pumps were so grossly oversized, the motor on Pump A
was not tripping out due to high load. Thus, the problem was not revealed
until the reliability engineer conducted his field test on Pump A. Operating
a pump such as this in the correct rotation but with the impeller reversed
would typically result in less flow and head than normal for a pump, and an
extremely low efficiency compared with the expected value, which is exactly
what the field test revealed.
The reliability engineer learned several very important lessons from this
experience. He established more stringent quality checks during the assembly of horizontal split case double suction pumps. Also, he instituted a practice of field testing all critical pumps as soon as they were put back in service
following an overhaul.
Similar cases have been found where a pump was operating in the opposite
direction of rotation than appropriate, or where an oversized impeller had
been put in a pump during an overhaul without first checking the impeller
trim against the pump data sheet. Again, both of these problems would have
been discovered much earlier if a quick field test were done at startup after
pump overhauls.
© 2008 Taylor & Francis Group, LLC
Case Studies
303
G. The Case of the Problematic Variable Speed Pump
1. Background
As part of a major upgrade being done to expand processing capacity,
Premier Paper Company needs to install a new boiler to produce additional
low pressure steam for steam injection, heating, and drying services. The
boiler will operate at 400 psi, and the feed water flow rate that is required to
the boiler will range between 500 and 1000 gpm, depending on process needs
and the plant’s production rate, which varies seasonally. The new boiler and
the associated boiler pump, piping, valves, and controls are being sized and
specified by the Premier Paper Company’s design engineering team.
Because of the variable nature of the Premier Paper Company’s production, many of the systems in the plant require variable flow rates. Through
the years, many of the pumps in these systems, such as the plant cooling
water pumps, have been refitted with variable frequency drives (VFDs) to
convert them from constant speed pumps to variable speed pumps. These
conversions have generated considerable energy savings, as well as eliminating control valves and their associated maintenance costs. Additionally, the
VFDs have permitted soft start of the pumps, which is favored by the plant’s
electrical engineering group. Chapter 6, Section IV discusses these and other
benefits that VFDs offer. Since the new boiler feed pump will have a variable
flow rate, the design engineering team plans to drive the boiler feed pump
through a VFD to achieve the energy and maintenance benefits that they see
with other pumps that have been retrofitted with VFDs. The plant operations
and maintenance groups have gained confidence in VFDs from their experience using them on other systems, and they recognize the savings in energy
and maintenance costs that the use of VFDs have produced, lending their
support to this design concept.
The design engineering team completed the sizing and selection of the
boiler feed system components. The pump was sized for a design flow rate
of 1000 gpm and a total head of 1000 ft. A four-stage pump was selected,
running at 1750 rpm. The piping for the boiler feed system was designed for
up to 2000 gpm, as the design engineering team anticipated that the needs
for low pressure steam in the plant will grow and that the boiler and pumps
eventually will have to be upgraded with higher capacity components. This
resulted in an expected velocity in the pipes that was lower than would normally be used for the planned 500- to 1000-gpm flow rate, but the design
engineering team determined that this would not have any ill effects on
plant operation, and they were willing to spend the additional money for the
larger than needed piping, because the pipe run, once built, would be quite
costly to replace due to details of the plant layout.
The NPSHa at the inlet of the boiler feed pump was calculated at 12 ft.
To have a pump with a low enough NPSH, the first stage was chosen with
the manufacturer with an enlarged eye area, to reduce the NPSHr for the
© 2008 Taylor & Francis Group, LLC
304
Pump Characteristics and Applications
pump to 7 ft. This resulted in a suction specific speed, Nss, of approximately
13,000 (per Equation 2.24, with Q of 1000, N of 1750, and NPSHr of 7). This
was higher than typically used in the plant, but the design engineering
team decided that since a VFD was being used, the boiler feed pump should
always operate near its BEP, hence the higher Nss pump selection.
The equipment, piping, fittings, and other system components were delivered to the plant and the boiler system was installed. The boiler system was
started up, and almost immediately, unforeseen problems arose. The boiler
feed pump operated fine at the upper flow rate (1000 gpm). But attempts to
slow the pump down to achieve the lower flow rate were met with extremely
high vibration levels on the pump and motor, causing the motor to trip. The
design engineering team was called in to troubleshoot the system.
2. Analysis of the Problem
Initial troubleshooting checks were done on the pump, confirming correct
rotation and good alignment of pump to driver. The NPSH for the system
was reviewed and found not to be a problem, as the pump has more than
adequate margin of NPSH at the highest running speed. At the lower speed,
the NPSHr of the pump was significantly lower and the NPSHa in the suction system was higher, so the NPSH margin was even higher than it was at
the higher speed. Consequently, NPSH was ruled out as the root cause of the
problem.
The next thing the design engineering team considered was whether
there might be a resonance problem with this pump at reduced speed. As
discussed in Chapter 4, Section XI, some variable speed pumps that are
somewhat long and flexible have a natural frequency that is below the maximum operating speed. Thus, when they are run at reduced speed, there is
the possibility that the operating speed might be too close to the natural
frequency of the pump. This can lead to a resonance problem, with its attendant high vibration levels. This problem is most likely to occur with long,
narrow pumps, such as vertical turbine and long horizontal split case multistage pumps. Although the boiler feed pump is a four-stage horizontal split
case pump, it appears to be fairly rigid and compact. The design engineering
team consulted with the pump manufacturer, which confirmed that the first
natural frequency of this pump was approximately 40% above the maximum
operating speed. This means the pump would never be run at a speed that
is any closer than 40% away from the natural frequency of the pump. At the
slower speed, the running speed would be even further away from the natural frequency of the pump. This ruled out resonance as a source of the pump
vibration problems at lower speeds.
The design engineering team finally decided that they must perform a full
hydraulic analysis of the boiler feed pump system, to see if that would lead
to the cause of the vibration problems at the lower pump speed. Part of that
analysis included generating the system head curve (see Chapter 2, Section IX
© 2008 Taylor & Francis Group, LLC
305
Case Studies
for discussion of how this curve is generated). The system head curve was
overlaid with the pump head–capacity curve at both the high and low running speeds and the cause of the vibration problem was revealed.
With the other systems that VFDs have been used on at the Premier Paper
plant thus far, most of the system head was composed of friction head. A
typical system head curve that is composed of mostly friction head (such as
a component cooling water system) is shown as System 1 in Figure 9.10, along
with a pump head–capacity curve at several operating speeds. The pump
iso-efficiency curves are also shown in Figure 9.10. As the pump is slowed
down, the iso-efficiency line of the pump nearly follows the System 1 system
head curve. The plotted line connecting the BEP point at various operating
speeds is called the iso-best-efficiency curve. This curve is nearly identical to
the location of the system head curve for the mostly friction System 1 system
head curve shown in Figure 9.10. This is the reason why, for most component
cooling water systems, as the pump is slowed down, the pump continues to
operate at its BEP. This concept is described in Chapter 6, Section IV, and was
shown in Figure 6.7.
However, in the case of the boiler feed system, the system head is not
mainly composed of friction head. On the contrary, the system head is mainly
composed of static head (the pressure being maintained in the boiler). The
only friction head in the system, the losses in the piping and valves leading
up to the boiler, is in fact even less than would normally be the case, due to
the fact that the piping system has been oversized to accommodate a future
higher flow rate. Thus, the system head curve starts at a high head (the boiler
1250
58%
Total head (ft)
1000
750
68% 73% 76%
77%
System 2
76%
73%
1750 rpm
1530 rpm
500
250
895 rpm
System 1
0
0
250
500
750
1000
1250
1500
1750
Flow rate (gpm)
FIGURE 9.10
System head curves for a system that is composed of mostly friction head (System 1) and for
a system that is composed of mostly static head (System 2), along with a pump head–capacity
curve showing one impeller diameter at several operating speeds, and the iso-efficiency curves.
© 2008 Taylor & Francis Group, LLC
306
Pump Characteristics and Applications
pressure) at zero flow, and the system head curve is a nearly flat curve due
to the relatively small amount of friction head in the system piping. This is
illustrated by the system head curve designated as System 2 in Figure 9.10.
When the pump operating in the system described by the System 2 curve
is slowed down from the maximum running speed, the operating point
quickly moves back well to the left of BEP, instead of continuing to operate
at or near BEP as occurs with a system that looks like the System 1 curve. As
the pump in System 2 is slowed down, it very quickly moves to a point on its
head–capacity curve where the flow is unstable due to suction recirculation.
The relatively high value of Nss for the boiler feed pump chosen means this
occurs at a higher percentage of BEP than otherwise might be expected. (See
Chapter 8, Section III.B.5 for a discussion of the narrower operating range
required on pumps with higher suction specific speed impellers.)
3. Solutions and Lessons Learned
The design engineering team came to an important conclusion, namely that
a VFD is not a good choice for a system in which a significant amount of
the total head is static head. Due to this situation, a relatively small drop in
pump speed moved the pump to a fairly low percentage of the BEP flow, in
this case approximately 43% of BEP. Since the chosen pump has a relatively
high Nss, the stable operating range of the pump was significantly reduced.
Operation at the slower speed moved the pump outside of the stable range,
producing recirculation, as described in Chapter 8, Section III.B.5, which
caused the excessive vibration that the pump exhibited.
There are several alternative approaches that would likely have produced
better results. The system should have been designed with higher NPSHa,
so that a pump with higher NPSHr, and thus lower Nss could have been
used. The design engineering team also could have considered a two-pump
scenario for the boiler feed design, operating the smaller pump at lower flow
ranges, the larger pump at intermediate ranges, and the two pumps in parallel at the highest flow rates.
H. The Case of the Shoe-Horned Wastewater Pumps
1. Background
The Pleasant Oaks Wastewater Utility (POWU) operates a 150 million gallon
per day wastewater treatment plant that was built in the 1970s. After many
years in operation, during which time the overall capacity of the treatment
plant was increased several times, the plant’s capacity had become maxed out.
The bottleneck in flow occurred at the channel that directed flow by gravity
from the primary to the secondary treatment processes within the plant.
The POWU engineering staff decided in 1995 to add a pump station in
the middle of the primary effluent channel to add additional head to the
© 2008 Taylor & Francis Group, LLC
Case Studies
307
wastewater stream, which would ultimately have the effect of adding nearly
20% to the operating capacity of the wastewater treatment plant, expanding
it to the full capacity for which it was permitted.
A new pump station, designated the Mid-Plant Pump Station (MPPS),
was designed and constructed by POWU in the primary effluent channel, at
the midpoint between the primary and secondary treatment processes. The
MPPS consisted of three submersible wastewater pumps, each rated at 60,000
gpm at 12 ft TH. The pump station was designed to have two pumps operating in parallel simultaneously, with the third pump serving as a standby.
When the POWU design engineering team reviewed the pump intake design
criteria found in the Hydraulic Institute (HI) during the design of the MPPS,
they saw that the physical limitations of the existing plant prohibited them
from meeting some of the HI recommended dimensions for the MPPS wet
well. For instance, they were not able to follow the HI recommendation with
regard to the distance of the MPPS pumps from each other, from the bottom
of the MPPS intake channel, and from the back wall, as well as the approach
distance and fluid velocity in the wet well as the wastewater approached the
MPPS pumps. The POWU engineering team realized that they were cutting
corners on the MPPS wet well, but hoped that the increased overall capacity of
the wastewater treatment plant would more than compensate for the fact that
the pumps had to be “shoehorned” into a tighter than optimum space.
From the day the MPPS was put into service, POWU experienced significant operations and maintenance headaches with the submersible wastewater pumps. The operation of the pumps was noisy, and the pumps and
motors experienced higher than expected vibration levels. Bearings in the
pump motors required frequent change-outs, and the mechanical seals
between the pumps and the submersible motors failed far too frequently,
several times causing intrusion of wastewater into the motors. In addition,
the pumps showed signs of fatigue loading on the impeller vanes, ultimately
leading to several broken impeller vanes and pump shafts. Air entrainment
was suspected of being a contributing factor to these failures, as some surface vortices were observed in the pump intakes. All too often, the pump and
motor failures would occur in the middle of stormy periods when they were
expected to run around the clock.
To compound the problems with too frequent repairs, the submersible
pumps did not have a proper crane and guide rail arrangement for pulling
and reinstalling the pumps, so temporary cranes had to be brought in every
time one of the pumps needed servicing.
There were other problems with the MPPS as well. No check or block valves
had been provided in the pump discharges (due to space limitations), resulting in the pumps bypassing flow backward through the standby pump, which
resulted in the pumps running further out on their curves than expected.
And, the pumps, which discharged directly into the secondary treatment
intake channel, created wave action and splash in this intake channel, which
caused health concerns from the excessive misting that this created.
© 2008 Taylor & Francis Group, LLC
308
Pump Characteristics and Applications
Ten years after the MPPS pumps were installed, POWU decided to replace the
submersible pumps with vertical column type axial flow pumps. Figure 9.11
shows a layout of the MPPS with the planned vertical axial flow pumps. This
change of pump style would help eliminate the problems caused by the inadequate lifting/guide rail assemblies for the submersible pumps, would allow
POWU to reorient the discharge pipes on the pumps to reduce splashing and
misting, and would enable them to add check valves in the pump discharges
(not shown in Figure 9.11). POWU decided to take this opportunity to look
more closely at the wet well in which the pumps were mounted, with the
hope that they could modify it to improve pump reliability.
2. Analysis of the Problem
Since POWU could not meet all of the requirements of the Hydraulic Institute,
they decided to seek technical help. They hired True Flow Hydraulic
Consultants, who recommended building of a scale model of the MPPS.
Although expensive, POWU decided that this investment was worthwhile,
given the disconnect between their physical space and ideal design and the
persistent maintenance problems.
The primary objective of the model study was to identify the adverse
conditions in the pump wet well and to develop modifications to the MPPS
structure that would produce acceptable flow characteristics and reduce
pump vibration, vortices, and air entrainment problems. Lack of uniformity
in flow approaching pumps in an open channel can result in excessive preswirl of flow approaching the impeller, free-surface and subsurface vortex
formation in the vicinity of the pump inlet, and nonuniform distribution
of flow approaching the impeller. Flow irregularities can, in turn, lead to
fatigue loads on the pump impeller and shaft, excessive vibration, and accelerated bearing and seal wear.
Since flow usually approaches the impellers axially, impellers generally
are designed with that in mind. Excessive swirl in the approach will cause
the flow to hit the impeller vanes at an angle, which can cause a drop in
pump head and capacity and a reduction in local pressure along the impeller
vane. In extreme cases, this pressure reduction can produce cavitation damage to the impeller, which exacerbates the vibration and fatigue being caused
by the air entrainment and uneven loading of the impellers.
Surface and subsurface vortices also can influence pump reliability and life.
These flow characteristics create a fluctuating load on the vanes of the impeller
as each vane passes through the low pressure vortex core. This leads to vibration, increased bearing wear, and fatigue loading. Surface vortices that are
capable of introducing entrained air into the pump can result in a reduction
of pump capacity, as well as additional nonuniform loading of the impeller.
The hydraulic model that was designed and built by True Flow Hydraulic
Consultants was a 1/8 scale model. For free-surface flow, such as associated
with a wet well, the gravitational and inertial forces are the dominant forces
© 2008 Taylor & Francis Group, LLC
309
Case Studies
Primary
treatment
effluent
channel
Sluice
gates
Secondary treatment
influent channel
Primary
treatment
effluent
channel
Sluice
gates
FIGURE 9.11
Planned layout of the midplant pump station in Case H.
© 2008 Taylor & Francis Group, LLC
310
Pump Characteristics and Applications
that define the flow in the wet well. As a result, the Froude number, as defined
below, is the key dimensionless variable that must be equal in the model and
prototype. That is.
Fr = v/(gl)1/2
(9.1)
where Fr is the Froude number, v is the velocity, g is the acceleration due to
gravity, and l is the characteristic length.
The model’s geometric 1/8 scale was selected to minimize viscous and surface tension scale effects. Also, the model had to be large enough to allow
engineers to see and accurately measure flow pre-swirl and velocities. The
model was constructed of wood framing, with acrylic plastic on some of the
walls to allow visual inspection of the flow as it approached the pumps in
the model. Flow was circulated through the model by a laboratory centrifugal pump, controlled by a flow valve located downstream of the laboratory
pump. The installation included swirl meters to measure the velocity and
angle of the pre-swirl, as well as dye injectors for a visual observation of
surface and subsurface vortices. The model included approximately 16 ft of
the primary effluent channel, the sluice gates that directed flow to the MPPS
pump wet well, the wet well including the three pump bays, the anti-vortex
cones beneath the pumps, the geometry of the pumps themselves, the discharge chamber into which the pump discharge was directed, and approximately 11 ft of the secondary treatment influent channel.
3. Solutions and Lessons Learned
The model was analyzed in several phases. It first was constructed to accurately represent the situation that existed at the MPPS, to establish a baseline. Next, proposed design modifications were introduced and tested. The
purpose of these modification tests was to develop and implement changes
to the design that would normalize the distribution and velocities of flow
approaching the pumps, reduce the amount of pre-swirl at the pumps, minimize the surface and subsurface vortex activity, and reduce the wave action
and splash in the discharge chamber. Altogether, more than 40 different
modifications were tested in the model study.
As a result of the model study, the following design changes were
implemented:
• Installed an angled baffle wall in each bay to decrease the formation
of surface vortices entering the pumps and to reduce flow pre-swirl;
• Raised the horizontal portion of the floor near the pumps 1.5 ft to
decrease the pump suction bell to floor clearance to control the formation of subsurface vortices;
• Installed side wall and back wall fillets and floor splitters beneath
each pump to minimize subsurface vortex activity; and
© 2008 Taylor & Francis Group, LLC
Case Studies
311
• Rotated the discharge chamber pump outlets by 25 degrees to
decrease the wave action and splash in the secondary treatment
influent channel.
The vertical column axial flow pumps were installed in the MPPS, which
had been modified as a result of the model study. POWU has completed
5 years of operation with this new arrangement, and the utility is quite
pleased with the MPPS operation with the above modifications. The pumps
run with notably less noise and vibration, and there is no visible evidence of
surface or subsurface vortices. Most importantly, the maintenance costs for
these pumps have been reduced considerably.
This case study illustrates the value, in certain situations, of a model study
for a wet well or intake structure, especially when physical limitations constrain the intake structure from being optimally designed. These model
studies are not inexpensive, but they may dramatically improve pump operation and reduce maintenance costs.
I. The Case of the High Suction Specific Speed Pump
1. Background
The Fluid Catalytic Cracking Unit (FCCU) for the ABC Oil Company’s Alpha
Refinery was originally designed in the late 1970s. As part of this FCCU
system, two single-stage end suction overhung raw gasoline pumps were
installed. Both pumps were driven by single-stage steam turbines. The
design flow of these pumps was 1250 gal/min at 750 ft total head, and the
design operating speed was 3700 rpm. Process requirements called for one
pump (Pump A) to be in operation and the other pump (Pump B) to serve as
a 100% spare, as is typical for refinery process pumps.
Approximately 20 years later, the FCCU went through an upgrade project, resulting in an increase in the capacity of the unit. This resulted in the
need to pump a higher capacity of raw gasoline. Rather than install two new
larger pumps, the project team decided to install only one new, larger pump
(Pump C), and to operate the two existing pumps (Pumps A and B) in parallel as the spare for this service.
The new Pump C is a double suction single stage between bearings pump,
driven by an electric motor at 3560 rpm. The design operating conditions for
this pump are 1510 gal/min at 825 ft total head.
2. Analysis of the Problem
Problems began almost immediately after the upgrade of the FCCU unit.
Operations soon learned that when they tried to operate the Pumps A and
B in parallel as a backup for the new Pump C, the mechanical seal in one or
both pumps would fail in short order. The failed pump would frequently be
© 2008 Taylor & Francis Group, LLC
312
Pump Characteristics and Applications
on fire when operations arrived on the scene. Operations began to refer to
Pumps A and B ironically as “fire pumps.” The problem was so severe that
operations would only run Pumps A and B when the new Pump C was out
of service for maintenance. Fortunately, this was not very often.
Operations had made several requests to have the two turbine-driven
pumps, Pumps A and B, replaced with a new turbine-driven pump that
would be identical to the larger electric motor-driven Pump C. Project engineering generated an estimate for this work, including taking into account
the lost production when Pump C was out of service, but this replacement
project seemed to always fall behind other more profitable or more urgent
projects. The machinery engineering group at the refinery knew that the
only way to resolve this problem was to find a solution that did not involve
replacement of these pumps. Given the penchant for Pumps A and B to fail
and cause fires when run in parallel, they also knew that they would have
to determine the cause of the failures without putting the pumps in service.
The curve and data sheets for Pumps A and B were examined by the
machinery engineering group, but the group failed to discover any obvious
problems. Figure 9.12 shows the performance curve for Pumps A and B, showing the single pump operation (the way the pumps were operated prior to the
upgrade) and the parallel operation of Pumps A and B as they were operated
after the upgrade. The pumps, when operating in parallel, were operating at
61% of the BEP, which normally is not a significant problem, as that is usually
well above the minimum continuous flow rate for process pumps in the refinery’s experience. The NPSHa exceeded the NPSHr by at least 15 ft.
The next step in the process was to determine the suction specific speed,
Nss, of Pumps A and B to see if that might provide a clue as to what might be
causing the problem. Equation 2.24 was used for this calculation, using the
BEP values of Q and NPSHr from the pump curve, and the design operating
speed of 3700 rpm. The impellers were very close to maximum diameter, on
which Equation 2.24 is based.
3. Solutions and Lessons Learned
Calculation of the suction specific speed, Nss, for Pumps A and B with Q of
1250 gpm and NPSHr of 9.5 ft identified a significant factor that was likely
contributing to the problem. Nss was calculated at 24,175. This number was
more than twice the value of Nss that the refinery would typically allow for a
service such as this. Per the discussion in Chapter 8, Section III, a pump with
this high a suction specific speed would likely have a very narrow stable
operating range, especially an older pump such as this one. Apparently the
original selection of Pumps A and B had been made with the thought that
the NPSHa would be quite a bit less than was actually available. The solution
at the time was a pump with low NPSHr, which meant that the pump chosen had a much higher suction specific speed, Nss, than was necessary. This
high value of Nss was not a problem while Pumps A and B were operating
© 2008 Taylor & Francis Group, LLC
313
Case Studies
1000
900
800
Two-pumps
in parallel
700
Total head (ft)
One-pump
operation
600
500
400
300
200
100
0
0
200
400
600
800
1000
1200
1400
1600
1800
Flow rate (gpm)
FIGURE 9.12
Performance curve for Pumps A and B in Case I, showing both one- and two-pump performance.
by themselves, as originally intended, because they were then operating
very near to the BEP, thus being in the stable operating window. However,
when the two pumps were operated in parallel after the FCCU upgrade, the
pumps were then operating at 61% of the BEP flow, outside the stable operating window.
The machinery engineering group decided at this point to explore the
option of performing a hydraulic re-rate on Pumps A and B. A hydraulic
re-rate usually consists of replacing the impeller with a different impeller
of a different specific speed, and in some cases modifying the pump volute
as well. Especially in the API 610 industry, where process pumps are both
very expensive and have a very long manufacturing lead time, a hydraulic
re-rate may often be the best overall solution, taking less time than buying
new equipment, and often saving money. Since the pump casing is not being
replaced, the hydraulic re-rate would not require any changes in the piping,
base plate, coupling, or driver for the pumps.
The machinery engineering group figured that an impeller would be
needed that would meet the required head and flow, but with a lower suction specific speed (e.g., a higher NPSHr) than the original Pumps A and B
impellers. Though they did not include it in their specification requirements
for the re-rate, they also wanted to find impellers that had a minimum 10%
rise from BEP to shutoff, which was their company’s standard for parallel
pump operation.
© 2008 Taylor & Francis Group, LLC
314
Pump Characteristics and Applications
They received proposals from three pump vendors. Vendor “A” was
selected, with an offering that had a suction specific speed that was 10,700
(the lowest offered), and the lowest cost proposal. They were also the only
supplier to meet the requirement for a 10% rise to shutoff from the design
point.
As part of the hydraulic re-rate, several modifications of the pump had to
be made. To meet the hydraulic requirements, the tongues on the volutes
were extended. It took several adjustments to get the volute configuration
set so that the desired hydraulic performance was produced. Also, the
case wear ring was replaced with a redesigned wear ring because the new
impeller has a smaller impeller eye. The new case wear ring provided a
smooth transition for the flow from the pump suction nozzle into the eye
of the new impeller.
One other factor could cause a significant reliability problem for this
equipment. When two pumps are operated in parallel, it is very important that they rotate at or very near the same speed. If they do not, the
two pumps will not share the flow evenly, and one pump will operate further back on its head capacity curve than the other. (This is discussed in
Chapter 2, Section X.) A pump operated in this manner is usually hydraulically unstable and will typically experience increased levels of vibration
and bearing loads. The life of the pump will be reduced with a failure
occurring in short order if the pump is operated too low below its best
efficiency point.
An examination of the two turbine drives for Pumps A and B revealed
that one turbine was equipped with a mechanical/hydraulic governor and
the other turbine was equipped with the original mechanical governor that
came on the turbine at the time of purchase. Neither turbine was equipped
with a tachometer. Achieving and maintaining the identical operating
speed for both pumps at the same rate would be very difficult to do without
tachometers and with the governors that were currently in service. The decision was made to install electronic governors on both turbines to ensure that
proper speed control was maintained. This would also address other reliability issues that were experienced with retrofitting mechanical/hydraulic
governors on these old turbines.
Although the total cost of this upgrade was not inexpensive, it was approximately one third the cost of the installation of a new pump and steam turbine (as estimated by the refinery’s Capital Project Organization). The option
of hydraulically re-rating these pumps also provided the ability to address
the problem in a timely fashion without requiring a unit shutdown to implement. Execution of this re-rate/upgrade should prevent any FCCU rate
reduction due to an outage of the large, motor-driven pump, and the associated negative financial impact on the company.
© 2008 Taylor & Francis Group, LLC
Appendix A: Major Suppliers of Pumps
in the United States by Product Type
Table A.1 provides a listing of most of the major pump suppliers in the
United States (both domestically manufactured and imported), segmented
according to the pump types they offer. This list is by no means exhaustive, containing only those manufacturers with which the author is familiar.
Readers who are interested in locating information on a particular pump
type should be able to use this appendix as a guide, and should be able to
locate manufacturers by a Web search or in a Thomas Register or similar index
of manufacturers. Note that the appendix lists the manufacturers alphabetically and that it includes both centrifugal and positive displacement pump
suppliers. The list is sorted by the major brand names by which the pump
products are most commonly known, and, where applicable, the parent company is shown following the brand name. The types of pumps manufactured
by each supplier are indicated in the table, with the product types referring
to the following key.
Product Type Key
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
Single-stage, single suction, close-coupled
Single-stage, single suction, frame-mounted
Self-priming centrifugal
Single-stage, double suction
Multistage
Submersible (except vertical turbine)
Vertical turbine (submersible)
Vertical turbine (lineshaft)
Mixed and axial flow
Sliding vane
Flexible impeller
Progressing cavity
Rotary external and internal gear
Lobe and circumferential piston
Multiple screw
Other rotary
Piston/plunger (nonmetering)
Diaphragm (nonmetering)
Metering
ANSI
API
315
A
X
X
X
X
X
X
X
X
X
X
X
X
X
-
Company
Abel Pumps, Roper Industries
ABS Pumps, Sulzer
A-C Pump, Xylem
Ace Pump Corporation
Aermotor Pumps, Pentair
Afton Pumps, Inc.
Aldrich, Flowserve
Alfa Laval
Allweiler Pumps, Colfax
All-Flo Pump Company
American Marsh
American Stainless Pumps
American Turbine Pump
Ampco Pumps
Ansimag, Inc., Sundyne
APV, SPX
Aquatec
Armstrong Pump
Aro Corporation, Flowserve
Ash Pumps, Weir
Aurora Pump Company, Pentair
Barber-Nichols
Barnes, Crane Pumps & Systems
Major Pump Suppliers
TABLE A.1
X
X
X
X
X
X
X
X
X
X
X
X
-
B
X
X
X
X
X
-
C
X
X
X
-
D
X
X
X
-
E
X
X
X
F
X
X
-
G
X
X
X
X
-
H
X
X
X
-
I
-
J
-
K
X
-
L
Product Type
-
M
X
X
X
X
-
N
X
-
O
X
-
P
X
X
X
-
Q
X
X
X
X
X
X
-
R
-
S
X
X
X
-
T
X
-
U
316
Appendix A
Bell & Gossett, Xylem
Berkeley, Pentair
Blackmer Pump, Dover
Blue White Industries
Bornemann, ITT
Bran + Luebbe, SPX
Bredel, Watson Marlow
Buffalo Pumps, Inc.
Burks, Crane Pumps & Systems
Byron Jackson, Flowserve
Camac Industries
Carver Pump Company
Cascade Pump Company
Caster Pumps, Sundyne
Cat Pumps Corporation
Chempump, Teikoku
Chicago Pump, Grundfos
ClydeUnion, SPX
Coffin Turbo Pumps
Corcoran Company
Corken, Idex
Cornell Pump, Roper Industries
CPC Internalift
Crisafulli Pump Company
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
-
X
X
X
-
X
X
X
X
-
X
X
X
X
X
X
X
X
-
X
X
X
X
X
X
-
X
X
-
X
X
X
X
X
-
-
X
-
X
-
-
X
-
X
X
-
X
X
X
-
X
X
-
X
X
X
-
X
X
X
X
X
(continued)
Appendix A
317
A
X
X
X
X
X
X
X
X
X
Company
Crown, Crane Pumps & Systems
Dean Pump, Met-Pro
Delasco, PCM
Deming, Crane Pumps & Systems
Dempster Industries
Dickow Pump Company
Diener Precision Pumps
Domestic, Xylem
Dorr Oliver, Inc., F.L. Smidth
Durco, Flowserve
Ebara
Edwards, Pentair
Elro
Envirotech, Weir
Essco Pump Division
Fairbanks Nijhuis
Filter Pump Industries
Finish Thompson
Flint & Walling, Zoeller
Flojet Corporation, Xylem
Floway Pump Company, Weir
Flotec, Pentair
Flowserve
Major Pump Suppliers
TABLE A.1 (Continued)
X
X
X
X
X
X
X
X
X
B
X
X
X
X
X
X
X
C
X
X
X
D
X
X
X
X
X
X
X
E
X
X
X
X
X
X
F
X
X
X
X
X
G
X
X
X
H
X
X
X
I
-
J
-
K
-
L
Product Type
X
X
-
M
X
N
-
O
X
X
X
X
-
P
X
-
Q
X
X
-
R
X
-
S
X
X
X
X
X
T
X
X
X
U
318
Appendix A
Fluid Metering, Inc.
Fluid-o-Tech
Flux Pumps
Flygt, ITT
FMC
Franklin Electric
Fristam
Fybroc Division, Met-Pro
Galigher, Weir
Gardner-Denver
Gator Pump
Geho, Weir
GIW Industries, KSB
Godwin Pumps, Xylem
Gorman-Rupp
Goulds Pumps, Inc., ITT
Goulds Water Technology, Xylem
Graco, Inc.
Granco Pumps
Graymills Corporation
Grindex
Griswold Pump Company
Grundfos Pumps Corporation
Gusher Pump
Hayward Tyler
Hazelton Pumps, Weir
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
-
X
X
-
X
X
X
X
X
X
-
X
X
X
X
X
X
X
X
X
X
X
X
X
-
X
X
X
-
X
X
X
-
X
X
-
-
-
X
-
X
X
-
-
X
X
-
X
X
X
X
X
-
X
X
X
X
-
X
X
-
X
X
X
X
(continued)
Appendix A
319
A
X
X
X
X
X
X
X
X
X
X
Company
HMD Kontro, Sundyne
HOMA Pump
Hotsy Corporation
Houttuin
Hydra-Cell, Wanner Engineering
Hydromatic, Pentair
Hypro Corporation, Pentair
IDP, Flowserve
IMO, Colfax
Innomag
Ikawi Pumps
Jabsco, Xylem
Jaeco Pump Company
Johnson, SPX
Johnston Pump, Sulzer Pumps
Klaus Union
Komline-Sanderson
Krebbs, F.L. Smitdth
Krogh Pump Company, Carver
KSB Pumps, Inc.
LaBour, Peerless, Grundfos
Laing, Xylem
Lancaster Pump, C-B Tool
Major Pump Suppliers
TABLE A.1 (Continued)
X
X
X
X
X
X
X
X
X
X
-
B
X
X
X
X
X
X
-
C
X
X
-
D
X
X
X
X
E
X
X
X
X
X
F
X
X
-
G
X
X
X
-
H
X
X
X
-
I
X
-
J
X
-
K
-
L
Product Type
X
X
-
M
X
X
-
N
X
X
X
-
O
X
X
-
P
X
X
X
X
X
-
Q
X
X
-
R
X
X
-
S
X
X
X
X
-
T
X
X
X
X
-
U
320
Appendix A
Layne Vertiline, Pentair
Leistritz Corporation
Lewa, Nikkiso
Lewis, Weir
Lincoln, SKF
Liquiflo, Picut Industries
Little Giant
LMI, Milton Roy
Lobee
Lowara, Xylem
Lutz Pumps, Inc.
Luwa
Maag Pump Systems, Textron
Magnatex
March Pumps
Marelli, Sundyne
Marlow Pumps, Xylem
Maso Sine Pump, Watson Marlow
McDonald A.Y.
McFarland, Tritan
Megator
Metso, Weir
Micropump, Idex
Milton Roy
Milton Roy, Hartell Div.
Morris, Grundfos
X
X
X
X
X
X
X
X
X
X
X
X
-
X
X
X
X
X
X
X
X
X
X
X
X
-
-
X
X
X
-
X
X
-
X
X
X
-
X
X
-
X
-
-
-
-
X
X
X
X
X
X
-
X
-
X
-
X
X
-
X
X
X
-
X
X
X
X
X
-
X
X
X
X
X
-
X
X
X
(continued)
Appendix A
321
A
X
X
X
X
X
X
X
X
X
X
X
X
X
-
Company
Moyno Inc.
MP Pumps
MTH
Multiflo, Weir
Multiquip
Myers, Pentair
Nagle Pumps, Inc.
National Pump, Gorman Rupp
Neptune, Roper Industries
Netzsch, Inc.
Nikkiso
Oberdorfer, Gardner Denver
Osmonics, G. E. Water Technology
Pacer Pumps, ASM Industries
Paco Pumps, Inc., Grundfos
Patterson Pump Company
PCM
Peerless Pump, Grundfos
Pioneer Pumps
Plenty, SPX
Price Pump Company
Procon Products
Major Pump Suppliers
TABLE A.1 (Continued)
X
X
X
X
X
X
X
X
X
X
-
B
X
X
X
X
X
X
X
X
-
C
X
X
X
-
D
X
X
X
X
X
-
E
X
X
X
X
X
X
X
-
F
X
X
-
G
X
X
X
-
H
X
X
-
I
X
X
J
X
-
K
X
X
X
X
-
L
Product Type
X
X
M
X
X
-
N
X
-
O
-
P
X
X
-
Q
X
X
-
R
X
X
-
S
X
-
T
-
U
322
Appendix A
ProMinent
Prosser, Crane Pumps & Systems
Pulsafeeder, Idex
Pumpex
Pumpworks 610
Reddy-Buffaloes Pump Company
Red Jacket Pumps, Xylem
Roper Pump, Roper Industries
Roth Pump Co.
Roto Jet, Weir
Rotor-Tech Inc.
Ruhrpumpen
Rule, Xylem
Schwing America
Scot Pump, Ardox
Seepex
Serfilco, Ltd.
Sethco Division, Met-Pro
Sherwood, Pentair
Shurflo, Pentair
Sier-Bath, Flowserve
Simflo Pumps, Inc.
SPP Pumps
Sta-Rite Industries, Pentair
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
-
X
X
X
X
X
X
X
X
X
-
X
X
X
X
X
X
X
X
X
X
-
X
X
X
X
X
X
X
X
X
X
-
X
-
X
X
-
X
X
-
X
X
X
X
-
-
X
-
X
X
X
-
X
-
X
-
X
X
X
-
X
(continued)
Appendix A
323
A
X
X
X
X
X
X
X
X
X
X
X
Company
Sulzer Pumps
Sundyne
Sunflo, Sundyne
Sykes, Ameripumps
Taco, Inc.
Tech-Mag
Teikoku
Texsteam, G.E. Energy
Tri-Rotor, Inc.
Tuthill Pump Company
United, Flowserve
Vanton Pump
Vaughn
Versa Matic, Idex
Vertiflo
Viking Pump, Inc., Idex
Vogel, Xylem
Warman, Weir
Major Pump Suppliers
TABLE A.1 (Continued)
X
X
X
X
X
X
X
X
X
X
B
X
X
X
X
X
-
C
X
X
-
D
E
X
X
X
X
F
-
G
X
X
-
H
I
X
X
X
X
-
J
-
K
-
L
Product Type
X
X
-
M
X
X
-
N
-
O
X
X
-
P
-
Q
X
-
R
X
-
S
X
-
T
X
X
X
-
U
324
Appendix A
Warren Pumps, Inc., Colfax
Warren Rupp, Inc., Idex
Watson Marlow
Waukesha Cherry Burrell, SPX
Wayne Water Systems, Scott Fetzer
Webster, Capital City Tool
Weil Pump Company
Weinman, Crane Pumps & Systems
Wemco, Weir
Wheatley Gaso, National Oilwell
Wilden, Dover
Wilfley & Sons, Inc.
Williams, Milton Roy
Wilo
Wilson-Snyder, Flowserve
Worthington, Flowserve
Yamada Pump
Yeomans, Grundfos
Zenith Pumps, Colfax
Zoeller Company
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
-
X
X
X
-
X
-
X
X
X
X
X
X
X
X
X
X
X
X
-
X
-
-
-
-
-
X
X
X
-
X
-
X
-
X
-
X
-
X
X
X
-
X
X
X
-
X
X
-
X
X
-
Appendix A
325
Appendix B: Conversion Formulae
The formulae used in this book are generally stated in U.S. Customary
System (USCS) units, the system most widely used by the pump industry
in the United States. This appendix provides simple conversion formulae for
USCS and SI (metric) units. The most common terms mentioned in this book
are stated in both units.
Multiply
Acres
Acres
Acres
Acres
Acre-feet
Acre-feet
Acre-feet
Atmospheres
Atmospheres
Atmospheres
Atmospheres
Atmospheres
Atmospheres
Barrels-oil
Barrels-beer
Barrels-whiskey
Barrels/Day-oil
Bags or sacks-cement
Board feet
British Thermal Units
British Thermal Units
British Thermal Units
British Thermal Units
British Thermal Units
B.T.U./min
B.T.U./min
B.T.U./min
B.T.U./min
Centares (Centiares)
Centigrams
Centiliters
Centimeters
By
43,560
4047
1.562 × 103
4840
43,560
325,851
1233.48
76.0
29.92
33.90
10,332
14.70
1.058
42
31
45
0.02917
94
144 sq in × 1 in
0.2520
777.6
3.927 × 104
107.5
To Obtain
Square feet
Square meters
Square miles
Square yards
Cubic feet
Gallons
Cubic meters
Cm of mercury
Inches of mercury
Feet of water
Kg/sq meter
Lb/sq in
Tons/sq ft
Gallons-oil
Gallons-beer
Gallons-whiskey
Gallons/min-oil
Pounds-cement
Cubic inches
Kilogram-calories
Foot-lb
Horsepower-hrs
2.928 × 104
Kilogram-meters
Kilowatt-hr
12.96
0.02356
0.01757
17.57
1
0.01
0.01
0.3937
Foot-lb/sec
Horsepower
Kilowatts
Watts
Square meters
Grams
Liters
Inches
(continued)
327
328
Appendix B
Multiply
Centimeters
Centimeters
Centimeters of mercury
Centimeters of mercury
Centimeters of mercury
Centimeters of mercury
Centimeters of mercury
Centimeters/sec
Centimeters/sec
Centimeters/sec
Centimeters/sec
Centimeters/sec
Centimeters/sec
Cms/sec/sec
Cubic centimeters
By
0.01
10
0.01316
0.4461
136.0
27.85
0.1934
1.969
0.03281
0.036
0.6
0.02237
3.728 × 10–4
0.03281
To Obtain
Meters
Millimeters
Atmospheres
Feet of water
Kg/sq meter
Lb/sq ft
Lb/sq in
Feet/min
Feet/sec
Kilometers/hr
Meters/min
Miles/hr
Miles/min
3.531 × 10–5
Feet/sec/sec
Cubic feet
Cubic centimeters
6.102 × 10–2
Cubic inches
Cubic centimeters
10–6
Cubic meters
Cubic centimeters
1.308 × 10–6
Cubic yards
Cubic centimeters
2.642 × 10–4
Gallons
Cubic centimeters
9.999 × 10–4
Liters
Cubic centimeters
2.113 × 10–3
Pints (liq.)
Cubic centimeters
1.057 × 10–3
Quarts (liq.)
Cubic feet
2.832 × 10–4
Cubic cms
Cubic feet
Cubic feet
Cubic feet
Cubic feet
Cubic feet
Cubic feet
Cubic feet
Cubic feet/min
Cubic feet/min
Cubic feet/min
Cubic feet/min
Cubic feet/sec
Cubic feet/sec
Cubic inches
Cubic inches
1728
0.02832
0.03704
7.48052
28.32
59.84
29.92
472.0
0.1247
0.4719
62.43
0.646317
448.831
16.39
5.787 × 10–4
Cubic inches
Cubic meters
Cubic yards
Gallons
Liters
Pints (liq.)
Quarts (liq.)
Cubic cms/sec
Gallons/sec
Liters/sec
Pounds of water/min
Millions gal/day
Gallons/min
Cubic centimeters
Cubic feet
Cubic inches
1.639 × 10–5
Cubic meters
Cubic inches
2.143 × 10–5
Cubic yards
Cubic inches
4.329 × 10–3
Gallons
(continued)
329
Appendix B
Multiply
By
To Obtain
Cubic inches
1.639 × 10–2
Liters
Cubic inches
Cubic inches
Cubic meters
0.03463
0.01732
Pints (liq.)
Quarts (liq.)
Cubic centimeters
Cubic meters
Cubic meters
Cubic meters
Cubic meters
Cubic meters
Cubic meters
Cubic meters
Cubic meters/hr
Cubic yards
Cubic yards
Cubic yards
Cubic yards
Cubic yards
Cubic yards
Cubic yards
Cubic yards
Cubic yards/min
Cubic yards/min
Cubic yards/min
Decigrams
Deciliters
Decimeters
Degrees (angle)
Degrees (angle)
Degrees (angle)
Degrees/sec
Degrees/sec
Degrees/sec
Dekagrams
Dekaliters
Dekameters
Drams
Drams
Drams
Fathoms
Feet
Feet
Feet
106
35.31
61023.7
1.308
264.2
999.97
2113
1057
4.40
764,554.86
27
46.656
0.7646
202.0
764.5
1616
807.9
0.45
3.366
12.74
0.1
0.1
0.1
60
0.01745
3600
0.01745
0.1667
0.002778
10
10
10
27.34375
0.0625
1.771845
6
30.48
12
0.3048
Cubic feet
Cubic inches
Cubic yards
Gallons
Liters
Pints (liq.)
Quarts (liq.)
Gallons/min
Cubic centimeters
Cubic feet
Cubic inches
Cubic meters
Gallons
Liters
Pints (liq.)
Quarts (liq.)
Cubic feet/sec
Gallons/sec
Liters/sec
Grams
Liters
Meters
Minutes
Radians
Seconds
Radians/sec
Revolutions/min
Revolutions/sec
Grams
Liters
Meters
Grains
Ounces
Grams
Feet
Centimeters
Inches
Meters
(continued)
330
Appendix B
Multiply
By
To Obtain
Feet
Feet of water
Feet of water
Feet of water
Feet of water
Feet of water
Feet/min
Feet/min
Feet/min
Feet/min
Feet/min
Feet/sec
Feet/sec
Feet/sec
Feet/sec
Feet/sec
Feet/sec
Feet/sec/sec
Feet/sec/sec
Foot-pounds
0.333
0.0295
0.8826
304.8
62.43
0.4335
0.5080
0.01667
0.01829
0.3048
0.01136
30.48
1.097
0.5924
18.29
0.6818
0.01136
30.48
0.3048
1.286 × 10–3
Yards
Atmospheres
Inches of mercury
Kg/sq meter
Lb/sq ft
Lb/sq inch
Centimeters/sec
Feet/sec
Kilometers/hr
Meters/min
Miles/hr
Centimeters/sec
Kilometers/hr
Knots
Meters/min
Miles/hr
Miles/min
Cms/sec/sec
Meters/sec/sec
British Thermal Units
Foot-pounds
5.050 × 10–7
Horsepower-hr
Foot-pounds
3.240 × 10–4
Kilogram-calories
Foot-pounds
Foot-pounds
0.1383
Kilogram-meters
Kilowatt-hours
Foot-pounds/min
2.140 × 10–5
0.01667
Foot-pounds/min
Foot-pounds/min
3.766 × 10–7
B.T.U./sec
3.030 × 10–5
Foot-pounds/sec
Horsepower
Foot-pounds/min
5.393 × 10–3
Gm-calories/sec
Foot-pounds/min
2.280 × 10–5
Kilowatts
Foot-pounds/sec
7.704 × 10–2
B.T.U./min
Foot-pounds/sec
1.818 × 10–3
Horsepower
Foot-pounds/sec
1.941 × 10–2
Kg-calories/min
Foot-pounds/sec
1.356 × 10–3
3785
0.1337
231
Kilowatts
Gallons
Gallons
Gallons
Gallons
Gallons
Gallons
Gallons
Gallons
3.785 × 10–3
4.951 × 10–3
3.785
8
4
Cubic centimeters
Cubic feet
Cubic inches
Cubic meters
Cubic yards
Liters
Pints (liq.)
Quarts (liq.)
(continued)
331
Appendix B
Multiply
Gallons-Imperial
Gallons-U.S.
Gallons water
Gallons/min
Gallons/min
Gallons/min
Grains (troy)
Grains (troy)
Grains (troy)
Grains/U.S. gal
Grains/U.S. gal
Grains/Imp. gal
Grams
Grams
Grams
Grams
Grams
Grams
Grams
Grams/cm
Grams/cu cm
Grams/cu cm
Grams/liter
Grams/liter
Grams/liter
Grams/liter
Hectares
Hectares
Hectograms
Hectoliters
Hectometers
Hectowatts
Horsepower
Horsepower
Horsepower
Horsepower
Horsepower
Horsepower
Horsepower
Horsepower (boiler)
Horsepower (boiler)
Horsepower-hours
By
1.20095
0.83267
8.345
2.228 × 10–3
0.06308
8.0208
0.06480
0.04167
2.0833 × 10–3
17.118
142.86
14.254
980.7
15.43
0.001
1000
0.03527
0.03215
2.205 × 10–3
5.600 ×10–3
62.43
0.03613
58.416
8.345
0.06242
1000
2.471
1.076 × 105
100
100
100
100
42.44
33,000
550
1.014
10.547
0.7457
745.7
33,493
9.809
2546
To Obtain
U.S. gallons
Imperial gallons
Pounds of water
Cubic feet/sec
Liters/sec
Cu ft/hr
Grams
Pennyweights (troy)
Ounces (troy)
Parts/million
Lbs/million gal
Parts/million
Dynes
Grains
Kilograms
Milligrams
Ounces
Ounces (troy)
Pounds
Pounds/inch
Pounds/cubic foot
Pounds/cubic inch
Grains/gal
Pounds/1000 gal
Pounds/cubic foot
Parts/million
Acres
Square feet
Grams
Liters
Meters
Watts
B.T.U./min
Foot-lb/min
Foot-lb/sec
Horsepower (metric)
Kg-calories/min
Kilowatts
Watts
B.T.U./hr
Kilowatts
B.T.U.
(continued)
332
Appendix B
Multiply
Horsepower-hours
Horsepower-hours
Horsepower-hours
Horsepower-hours
Inches
Inches of mercury
Inches of mercury
Inches of mercury
Inches of mercury
Inches of mercury (32°F)
Inches of water
Inches of water
Inches of water
Inches of water
Inches of water
Inches of water
Kilograms
Kilograms
Kilograms
Kilograms
Kilograms-cal/sec
Kilograms-cal/sec
Kilograms-cal/sec
Kilograms-cal/sec
Kilogram-cal/min
Kilogram-cal/min
Kilogram-cal/min
Kgs/meter
Kgs/sq meter
Kgs/sq meter
Kgs/sq meter
Kgs/sq meter
Kgs/sq meter
Kgs/sq millimeter
Kiloliters
Kilometers
Kilometers
Kilometers
Kilometers
Kilometers
Kilometers/hr
Kilometers/hr
By
To Obtain
1.98 × 10
641.6
2.737 × 105
0.7457
2.540
0.03342
1.133
345.3
70.73
0.491
0.002458
0.07355
25.40
0.578
5.202
0.03613
980,665
2.205
1.102 × 10–3
103
3.968
3086
5.6145
4186.7
3085.9
0.09351
69.733
0.6720
9.678 × 10–5
3.281 × 10–3
2.896 × 10–3
0.2048
1.422 × 10–3
106
103
105
3281
103
0.6214
1094
27.78
54.68
Foot-lb
Kilogram-calories
Kilogram-meters
Kilowatt-hours
Centimeters
Atmospheres
Feet of water
Kg/sq meter
Lb/sq ft
Lb/sq inch
Atmospheres
Inches of mercury
Kg/sq meter
Ounces/sq inch
Lb/sq foot
Lb/sq inch
Dynes
Lb
Tons (short)
Grams
B.T.U./sec
Foot-lb/sec
Horsepower
Watts
Foot-lb/min
Horsepower
Watts
Lb/foot
Atmospheres
Feet of water
Inches of mercury
Lb/sq foot
Lb/sq inch
Kg/sq meter
Liters
Centimeters
Feet
Meters
Miles
Yards
Centimeters/sec
Feet/min
6
(continued)
333
Appendix B
Multiply
Kilometers/hr
Kilometers/hr
Kilometers/hr
Kilometers/hr
Km/hr/sec
Km/hr/sec
Km/hr/sec
Kilowatts
Kilowatts
Kilowatts
Kilowatts
Kilowatts
Kilowatts
Kilowatt-hours
Kilowatt-hours
Kilowatt-hours
Kilowatt-hours
Kilowatt-hours
Liters
Liters
Liters
Liters
Liters
Liters
Liters
Liters
Liters/min
Liters/min
Lumber width (in.) ×
thickness (in.)/12
Meters
Meters
Meters
Meters
Meters
Meters
Meters/min
Meters/min
Meters/min
Meters/min
Meters/min
Meters/sec
Meters/sec
By
To Obtain
0.9113
0.5399
16.67
0.6214
27.78
0.9113
0.2778
56.907
4.425 × 104
737.6
1.341
14.34
103
3414.4
2.655 × 106
1.341
860.4
3.671 × 105
103
0.03531
61.02
10–3
1.308 × 10–3
0.2642
2.113
1.057
5.886 × 10–4
4.403 × 10–3
Length (ft)
Feet/sec
Knots
Meters/min
Miles/hr
Cm/sec/sec
Ft/sec/sec
Meters/sec/sec
B.T.U./min
Foot-lbs/min
Foot-lbs/sec
Horsepower
Kg-calories/min
Watts
B.T.U.
Foot-lb
Horsepower-hr
Kilogram-calories
Kilogram-meters
Cubic centimeters
Cubic feet
Cubic inches
Cubic meters
Cubic yards
Gallons
Pints (liq.)
Quarts (liq.)
Cubic ft/sec
Gal/sec
Board feet
100
3.281
39.37
10–3
10
1.094
1.667
3.281
0.05468
0.06
0.03728
196.8
3.281
Centimeters
Feet
Inches
Kilometers
Millimeters
Yards
Centimeters/sec
Feet/min
Feet/sec
Kilometers/hr
Miles/hr
Feet/min
Feet/sec
(continued)
334
Appendix B
Multiply
Meters/sec
Meters/sec
Meters/sec
Meters/sec
Microns
Miles
Miles
Miles
Miles
Miles/hr
Miles/hr
Miles/hr
Miles/hr
Miles/hr
Miles/hr
Miles/min
Miles/min
Miles/min
Miles/min
Milliers
Milligrams
Millimeters
Millimeters
Millimeters
Milligrams/liter
Million gal/day
Miner’s inches
Minutes (angle)
Ounces
Ounces
Ounces
Ounces
Ounces
Ounces
Ounces
Ounces (troy)
Ounces (troy)
Ounces (troy)
Ounces (troy)
Ounces (troy)
Ounces (fluid)
Ounces (fluid)
By
3.6
0.06
2.287
0.03728
10–6
1.609 × 105
5280
1.609
1760
44.70
88
1.467
1.609
0.8689
26.82
2682
88
1.609
60
103
10–3
10–3
0.1
0.03937
1
1.54723
1.5
2.909 × 10–4
16
437.5
0.0625
28.3495
0.9115
2.790 × 10–5
2.835 × 10–5
480
20
0.08333
31.10348
1.09714
1.805
0.02957
To Obtain
Kilometers/hr
Kilometers/min
Miles/hr
Miles/min
Meters
Centimeters
Feet
Kilometers
Yards
Centimeters/sec
Feet/min
Feet/sec
Kilometers/hr
Knots
Meters/min
Centimeters/sec
Feet/sec
Kilometers/min
Miles/hr
Kilograms
Grams
Liters
Centimeters
Inches
Parts/million
Cubic ft/sec
Cubic ft/min
Radians
Drams
Grains
Pounds
Grams
Ounces (troy)
Tons (long)
Tons (metric)
Grains
Pennyweights (troy)
Pounds (troy)
Grams
Ounces (avoir.)
Cubic inches
Liters
(continued)
335
Appendix B
Multiply
Ounces/sq inch
Parts/million
Parts/million
Parts/million
Pennyweights (troy)
Pennyweights (troy)
Pennyweights (troy)
Pennyweights (troy)
Pounds
Pounds
Pounds
Pounds
Pounds
Pounds
Pounds
Pounds (troy)
Pounds (troy)
Pounds (troy)
Pounds (troy)
Pounds (troy)
Pounds (troy)
Pounds (troy)
Pounds (troy)
Pounds (troy)
Pounds of water
Pounds of water
Pounds of water
Pounds of water/min
Pounds/cubic foot
Pounds/cubic foot
Pounds/cubic foot
Pounds/cubic inch
Pounds/cubic inch
Pounds/cubic inch
Pounds/foot
Pounds/inch
Pounds/sq foot
Pounds/sq foot
Pounds/sq foot
Pounds/sq inch
Pounds/sq inch
Pounds/sq inch
By
0.0625
0.0584
0.07015
8.345
24
1.55517
0.05
4.1667 × 10–3
16
256
7000
0.0005
453.5924
1.21528
14.5833
5760
240
12
373.2417
0.822857
13.1657
3.6735 × 10–4
4.1143 × 10–4
3.7324 × 10–4
0.01602
27.68
0.1198
2.670 × 10–4
0.01602
16.02
5.787 × 10–4
27.68
2.768 × 104
1728
1.488
1152
0.01602
4.882
6.944 × 10–3
0.06804
2.307
2.036
To Obtain
Lb/sq inch
Grains/U.S. gal
Grains/Imp. gal
Lbs/million gal
Grains
Grams
Ounces (troy)
Pounds (troy)
Ounces
Drams
Grains
Tons (short)
Grams
Pounds (troy)
Ounces (troy)
Grains
Pennyweights (troy)
Ounces (troy)
Grams
Pounds (avoir.)
Ounces (avoir.)
Tons (long)
Tons (short)
Tons (metric)
Cubic feet
Cubic inches
Gallons
Cubic ft/sec
Grams/cubic cm
Kgs/cubic meters
Lbs/cubic inch
Grams/cubic cm
Kg/cubic meter
Lb/cubic foot
Kg/meter
Grams/cm
Feet of water
Kg/sq meter
Pounds/sq inch
Atmospheres
Feet of water
Inches of mercury
(continued)
336
Appendix B
Multiply
Pounds/sq inch
Quadrants (angle)
Quadrants (angle)
Quadrants (angle)
Quarts (dry)
Quarts (liq.)
Quintal, Argentine
Quintal, Brazil
Quintal, Castile, Peru
Quintal, Chile
Quintal, Mexico
Quintal, Metric
Quires
Radians
Radians
Radians
Radians/sec
Radians/sec
Radians/sec
Radians/sec/sec
Radians/sec/sec
Reams
Revolutions
Revolutions
Revolutions
Revolutions/min
Revolutions/min
Revolutions/min
Revolutions/min/min
Revolutions/min/min
Revolutions/sec
Revolutions/sec
Revolutions/sec
Revolutions/sec/sec
Revolutions/sec/sec
Seconds (angle)
Square centimeters
Square centimeters
Square centimeters
Square centimeters
Square feet
Square feet
By
703.1
90
5400
1.571
67.20
57.75
101.28
129.54
101.43
101.41
101.47
220.46
25
57.30
3438
0.637
57.30
0.1592
9.549
573.0
0.1592
500
360
4
6.283
6
0.1047
0.01667
1.745 × 10–3
2.778 × 10–4
360
6.283
60
6.283
3600
4.848 × 10–6
1.076 × 10–3
0.1550
10–4
100
2.296 × 10–5
929.0
To Obtain
Kgs/sq meter
Degrees
Minutes
Radians
Cubic inches
Cubic inches
Pounds
Pounds
Pounds
Pounds
Pounds
Pounds
Sheets
Degrees
Minutes
Quadrants
Degrees/sec
Revolutions/sec
Revolutions/min
Revs/min/min
Revs/sec/sec
Sheets
Degrees
Quadrants
Radians
Degrees/sec
Radians/sec
Revolutions/sec
Rads/sec/sec
Revs/sec/sec
Degrees/sec
Radians/sec
Revolutions/min
Radians/sec/sec
Revs/min/min
Radians
Square feet
Square inches
Square meters
Square millimeters
Acres
Square centimeters
(continued)
337
Appendix B
Multiply
Square feet
Square feet
Square feet
Square feet
1/Sq ft/gal/min
Square inches
Square inches
Square inches
Square kilometers
Square kilometers
Square kilometers
Square kilometers
Square kilometers
Square meters
Square meters
Square meters
Square meters
Square miles
Square miles
Square miles
Square miles
Square millimeters
Square millimeters
Square yards
Square yards
Square yards
Square yards
Temp (°C) +273
Temp (°C) +17.78
Temp (°F) +460
Temp (°F) –32
Tons (long)
Tons (long)
Tons (long)
Tons (metric)
Tons (metric)
Tons (short)
Tons (short)
Tons (short)
Tons (short)
Tons (short)
Tons (short)
By
144
0.09290
3.587 × 10–4
0.111
8.0208
6.452
6.944 × 10–3
645.2
247.1
10.76 × 106
106
0.3861
1.196 × 106
2.471 × 10–4
10.76
3.861 × 10–7
1.196
640
27.88 × 106
2.590
3.098 × 106
0.01
1.550 × 10–3
2.066 × 10–4
9
0.8361
3.228 × 10–7
1
1.8
1
5/9
1016
2240
1.12000
103
2205
2000
32,000
907.1843
2430.56
0.89287
29166.66
To Obtain
Square inches
Square meters
Square miles
Square yards
Overflow rate (ft/hr)
Square centimeters
Square feet
Square millimeters
Acres
Square feet
Square meters
Square miles
Square yards
Acres
Square feet
Square miles
Square yards
Acres
Square feet
Square kilometers
Square yards
Square centimeters
Square inches
Acres
Square feet
Square meters
Square miles
Abs temp (°C)
Temp (°F)
Abs. temp (°F)
Temp (°C)
Kilograms
Pounds
Tons (short)
Kilograms
Pounds
Pounds
Ounces
Kilograms
Pounds (troy)
Tons (long)
Ounces (troy)
(continued)
338
Appendix B
Multiply
Tons (short)
Tons of water/24 hrs
Tons of water/24 hrs
Tons of water/24 hrs
Watts
Watts
Watts
Watts
Watts
Watts
Watt-hours
Watt-hours
Watt-hours
Watt-hours
Watt-hours
Watt-hours
Yards
Yards
Yards
Yards
By
To Obtain
0.90718
83.333
0.16643
1.3349
0.05686
44.25
0.7376
1.341 × 10
0.01434
10–3
3.414
2655
1.341 × 10–3
0.8604
367.1
10–3
91.44
3
36
0.9144
Tons (metric)
Pounds water/hr
Gallons/min
Cu ft/hr
B.T.U./min
Foot-lbs/min
Foot-lbs/sec
Horsepower
Kg-calories/min
Kilowatts
B.T.U.
Foot-lbs
Horsepower-hrs
Kilogram-calories
Kilogram-meters
Kilowatt-hours
Centimeters
Feet
Inches
Meters
References
1. Hardee, R.T. and Sines, J.L., Piping System Fundamentals, 2nd Edition, Engineered
Software, Lacey, WA, 2009.
2. Hydraulic Institute, Engineering Data Book, 2nd Edition. Parsippany, NJ, 1990.
3. Karassik, I., Messina, J., Cooper, P., and Heald, C., Pump Handbook, 4th Edition,
McGraw-Hill, New York, 2007.
4. Hydraulic Institute Standards, Hydraulic Institute, Parsippany, NJ, multiple standards, various dates.
5. Stepanoff, A.J., Centrifugal and Axial Flow Pumps, 2nd Edition, Krieger Publishing,
Melbourne, FL, 1992.
6. Lobanoff, V. and Ross, R.R., Centrifugal Pumps: Design and Application, 2nd
Edition, Gulf Publishing, Houston, TX, 1992.
7. Chen, C.C., “Cope with Dissolved Gases in Pump Calculations,” Chemical
Processing, October 1993.
8. API Standard 610, Centrifugal Pumps for Petroleum, Petrochemical, and Natural Gas
Industries, 11th Edition, American Petroleum Institute, Washington, DC, 2010.
9. ASME B73.1 (ANSI B73.1), Specification for Horizontal End Suction Centrifugal
Pumps for Chemical Process, American National Standards Institute, Washington,
DC, 2001 (R2007).
10. Piotrowski, J., Shaft Alignment Handbook, 3rd Edition, CRC Press, Boca Raton,
FL, 2006.
339
Mechanical Engineering
Pump
Characteristics
and Applications
THIRD EDITION
Michael Volk
Providing a wealth of information on pumps and pump systems, Pump Characteristics
and Applications, Third Edition details how pump equipment is selected, sized, operated, maintained, and repaired. The book identifies the key components of pumps and
pump accessories, introduces the basics of pump and system hydraulics as well as more
advanced hydraulic topics, and details various pump types, as well as special materials
on seals, motors, variable frequency drives, and other pump-related subjects. It uses example problems throughout the text, reinforcing the practical application of the formulae
and analytical presentations. It also includes new images highlighting the latest generation of pumps and other components, explores troubleshooting options, and incorporates
relevant additions into the existing chapters.
What’s NeW iN this editioN:
• Includes more than 150 full-color images that significantly improve
the reader’s ability to understand pump drawings and curves
• Introduces a new chapter on pump case studies in a format that provides
case study background, analysis, solutions, and lessons learned
• Presents important new updates and additions to other chapters
• Includes a ten-step procedure for determining total pump head
• Discusses allowable and preferred operating ranges for centrifugal pumps
• Provides charts covering maximum and normally attainable pump efficiencies,
performance corrections for slurry pumps, and mechanical seal flush plans
Pump Characteristics and Applications, Third Edition is appropriate for readers with all
levels of technical experience, including engineering and pump industry professionals,
pump operators and maintenance technicians, upper-level undergraduate and graduate
students in mechanical engineering, and students in engineering technology programs.
K15984
an informa business
6000 Broken Sound Parkway, NW
Suite 300, Boca Raton, FL 33487
711 Third Avenue
New York, NY 10017
2 Park Square, Milton Park
Abingdon, Oxon OX14 4RN, UK